Beruflich Dokumente
Kultur Dokumente
George Symeonidis
All rights reserved. No part of this book may be reproduced in any form by any elec-
tronic or mechanical means (including photocopying, recording, or information storage
and retrieval) without permission in writing from the publisher.
This book was set in Palatino in ‘3B2’ by Asco Typesetters, Hong Kong and was
printed and bound in the United States of America.
Symeonidis, George.
The effects of competition : cartel policy and the evolution of strategy and structure in
British industry / George Symeonidis.
p. cm.
Includes bibliographical references and index.
ISBN 0-262-19468-6 (hc. : alk. paper)
1. Corporations—Great Britain—History—20th century. 2. Cartels—Great
Britain—History—20th century. 3. Strategic alliances (Business)—Great
Britain—History—20th century. 4. Competition—Great Britain—History—20th
century. 5. Big business—Great Britain—History—20th century. 6. Industrial policy
—Great Britain—History—20th century. 7. Prices—Great Britain—History—20th
century. 8. Costs, Industrial—Great Britain—History—20th century. I. Title.
HD2845 .S95 2002
338.8 0 7—dc21 2001044444
Contents
Acknowledgments ix
1 An Introductory Overview 1
1.1 A Natural Experiment 1
1.2 Key Theoretical Concepts 3
1.3 Empirical Methodology 7
1.4 Outline of the Book 12
1.5 Policy Implications 18
Epilogue 327
The research for this book began several years ago, when I was a
graduate student at the London School of Economics, and continued
after I took my current post at the University of Essex. My intellec-
tual debt to John Sutton is almost too obvious to be acknowledged.
His invaluable advice in the course of many and long discussions
has profoundly influenced my work, and I am also grateful to him for
numerous insightful comments on various drafts, including those on
the penultimate draft of this book.
I have also benefited greatly from discussions with several other
people on various aspects of the research presented here. I am espe-
cially obliged to V. Bhaskar, Paul Geroski, Tim Hatton, Sajal Lahiri,
Stephen Machin, Pedro Marin, Mark Schankerman, Philip Vergau-
wen, Michael Waterson, and Hugh Wills for their very helpful advice
and suggestions at various stages of the project. I also wish to
thank seminar audiences at the London Business School; the London
School of Economics; Nuffield College (Oxford); the Universities of
Athens, Carlos III (Madrid), East Anglia, Essex, Leuven, Maastricht,
and Warwick; and at numerous conferences, where parts of this re-
search were presented over the past few years.
I owe a very particular debt to David Genesove and Stephen
Martin, both of whom read the entire manuscript and provided very
detailed and thoughtful comments. Their suggestions led to signifi-
cant improvements in the final version.
The financial assistance of the Greek State Scholarships Founda-
tion helped me begin the research reported in this book. Some of this
work has appeared in different form in a number of journal articles,
and I wish to thank the publishers of Economica (Blackwell Publishers
and London School of Economics), the Journal of Industrial Economics
(Blackwell Publishers), and the Journal of Economics and Manage-
x
ment Strategy (MIT Press) for permission to use material from these
sources. I am indebted to Terry Vaughn, former economics editor at
the MIT Press, for his help and encouragement, and I am also thankful
to John Covell for dealing very efficiently with the manuscript.
Finally, my warmest thanks go to my parents, Antonis and Niki,
and to my wife, Angelina, for their continuous support and encour-
agement over many years.
1 An Introductory Overview
costs are sunk before firms set their price or output, and they are
endogenous because they can vary depending on each firm’s own
decisions.
It is, of course, true that firms may make decisions on nonprice
variables in all industries. The concept of an exogenous sunk cost
industry is simply a useful polar case that captures the fact that
in many industries, advertising and R&D are not important stra-
tegic considerations. In other words, the general theoretical model
throughout this book is one that allows for endogenous sunk costs.
But within that model there is a special limiting case which arises
when advertising effectiveness and technological opportunity are
both very low. In this limiting case, firms spend nothing on adver-
tising or R&D. Thus the equilibria of this limiting-case model coin-
cide with the equilibria of an elementary model in which there is no
advertising or R&D, and hence the only sunk cost is the exogenously
determined cost of entry. The elementary model can be used to ana-
lyze industries with zero or very low advertising and R&D expendi-
ture. On the other hand, the general model is the appropriate model
for advertising-intensive and R&D-intensive industries.
The notion that certain costs are sunk is closely related to the dis-
tinction between firms’ long-run and short-run decisions. At any
point in time, some variables are more difficult to change than
others. For instance, the decision to enter an industry by setting up a
plant is more difficult to reverse, once implemented, than the choice
of a price for the product. The reason is that capital stock is difficult
to resell and depreciates only gradually, while price setting does not
involve any kind of irreversible investment. It follows that the entry
decision must be taken as given, and the cost of setting up a plant is
largely a sunk cost, when a firm sets a price for its product. A game
structure that captures this distinction is that of the multistage game,
with the long-run decisions being taken at the earlier stages and
regarded as given when the short-run decisions are taken in the late
stages of the game.
The simplest structure that can be used to model competition in
an exogenous sunk cost industry is a two-stage game where firms
decide whether to enter or not to enter at stage 1, then choose a price
or a level of output at stage 2. To find the equilibrium of such a
game, we have to ask how the entry decision is influenced by the
anticipated outcome of short-run price or output competition. Market
structure, that is, the number of firms that enter the market and their
An Introductory Overview 7
The construction of the data set for this book has involved examining
a very large number of restrictive agreements registered under the
1956 Restrictive Trade Practices Act. Other sources of information
on competition in British manufacturing industry during the 1950s,
1960s, and 1970s were also used. It has therefore been possible to
assign all manufacturing industries to one of three groups: those
affected by the legislation, those not affected, and those which, for
one or the other reason, were difficult to classify. The competition
data were matched against data on several other variables, including
concentration, firm and plant numbers, profits and profit margins,
advertising intensity, and innovation counts, obtained from a variety
of official and unofficial sources. Several different samples of indus-
tries were constructed and used to analyze a variety of empirical
issues. The ambiguous group of industries was excluded from these
samples, and the results were based on comparisons between the
group of industries with a change in competition regime and the
group without such a change.
The time period considered spans a total of more than twenty
years, from the mid-1950s to the mid-1970s. For many of the vari-
ables, data are available only at roughly five-year intervals during
most of this period. In particular, Census of Production data are avail-
able for 1954, 1958, 1963, 1968, and then annually from 1970 onward.
The years 1954 and 1958 are ‘‘before’’ dates in the natural experiment,
the former because the legislation had not yet been introduced, the
8 Chapter 1
Two further remarks on the present research design and the links
between theory and empirical testing are in order. Much of the
theoretical and empirical analysis in this book abstracts from one
potentially important factor, namely, the fact that more intense price
competition may affect firms’ incentives to reduce their variable costs.
This can take the form of increased effort by workers and managers,
a relative reduction in wages and salaries, or the adoption of a better
technology.
While these are important issues, they are outside the scope of
this book. The assumption made here is that any such reduction in
variable costs will be largely passed on to prices and will have little
effect on price-cost margins. On the other hand, changes in price-
cost margins will be largely driven by changes in competition, advertis-
ing, and R&D intensity, and the number of firms. Consequently, any
potential effects of competition on the labor market and on produc-
tivity will not substantially alter the basic links between competition,
on the one hand, and market structure, advertising intensity, innova-
tion, and profitability, on the other, which are the subject of this book.
It may also be asked whether the assumption that firms are sym-
metric and produce one product each, which is made in the theoret-
An Introductory Overview 11
turing industry from the early 1950s to the mid-1970s. The purpose
of this chapter is to provide the essential background information on
the natural experiment analyzed in the book. I therefore begin by
describing the structure of the 1956 legislation and its implementa-
tion. I explain, among other things, why the large majority of indus-
tries registered their restrictive agreements rather than immediately
dropping them or secretly continuing them, and why it took more
than a decade for the effects of the 1956 Act on competition in British
industry to be fully realized. Moreover, rather than assume that the
Act had a significant effect on firm conduct, I set out to justify this
claim on the basis of the available case-study evidence. The conclu-
sion is that price competition did indeed intensify, following the
introduction of restrictive practices legislation, in the large majority
of the previously cartelized industries.
A more in-depth analysis of competition and collusion in the 1950s
is carried out in chapter 3. In the first part of this chapter I look at
the state of competition across British manufacturing industries in the
1950s, and I argue that the existing evidence allows us to distinguish
three groups: (1) a group of collusive industries (defined as indus-
tries with explicit agreements significantly restricting competition),
the large majority of which subsequently experienced a change of
competition regime; (2) a group of industries without explicit collu-
sive agreements, which therefore experienced no change in regime;
and (3) a residual group of industries for which we are uncertain as
to their state of competition. I also describe in some detail the con-
struction of the competition data used throughout the book.
In the second part of the chapter I examine which structural in-
dustry characteristics were associated with the occurrence of collu-
sion in the UK in the 1950s. The purpose of this exercise is twofold.
On the one hand, it is intended as a contribution to the literature on
the factors facilitating or hindering collusion. The significance of this
question is rather obvious: attempts by firms to establish price-fixing
arrangements are a matter of great concern for public policy, since
such behavior is regarded as being against the public interest, and is
therefore illegal under most competition laws. On the other hand, the
final part of chapter 3 prepares the way for the analysis of the natural
experiment by examining the similarities as well as the differences in
initial conditions between the two groups of industries. A key result
is that the two groups were broadly similar with respect to initial
market structure, but rather different with respect to capital intensity
14 Chapter 1
Price Competition and Profitability: Are Cartel Laws Bad for Business?
What the results of this book suggest is that the level of excess
profits does not depend, in the long run, on firms’ pricing conduct,
because of forces such as entry and exit that push industries toward
the zero-profit equilibrium. It would be misleading to interpret this as
suggesting that competition policy has set itself a relatively easy task
because, by and large, free entry eliminates excess profits in the long
run. Rather, by emphasizing the key role of free entry in restraining
higher than normal profits, this book suggests that, in addition to the
monitoring of firm conduct, the monitoring of entry conditions into
industries is a key priority for competition policy.
2 Cartel Policy and the
Evolution of Competition
in British Industry
2.1 Introduction
1. Detailed information on the structure of the 1948 Act, the role and the reports of the
Monopolies and Restrictive Practices Commission during the 1950s, and the impact of
these reports is contained in Guenault and Jackson (1974), Rowley (1966), Swann et al.
(1974), and Mercer (1995), among others.
Cartel Policy and Competition in British Industry 23
2. Aggregated rebates are discounts to distributors on the basis of the total quantity
purchased from association firms. Collective exclusive dealing is the practice whereby
distributors are required to deal only with association firms if they are to be supplied
on best terms or at all.
24 Chapter 2
3. Hunter (1966), Swann et al. (1974), and Wilberforce et al. (1966) describe in detail
the structure and the implementation of the 1956 Act. Subsequent developments in UK
cartel legislation until the early 1970s are discussed in Swann et al. (1974) and also in
the various reports of the Registrar of Restrictive Trading Agreements (RRTA, 1961–
1973). The reports of the Registrar also provide details on the implementation of the
legislation over the period 1956–1972, including brief accounts of cases examined by
the Restrictive Practices Court.
Cartel Policy and Competition in British Industry 25
little prepared for the strict attitude of the Restrictive Practices Court,
which eventually led to the abolition of cartels in Britain.
There is little doubt that the (Conservative) government had to
accommodate a lot of pressure from interested parties, especially
business, during the period of the drafting of the Bill. Hunter (1966,
pp. 85–86) argues that the overall structure of the 1956 Act was a
compromise between business interests, public opinion, and the gov-
ernment’s eagerness to promote economic liberalism. For instance,
he sees the provisions of the Act for more effective enforcement of
individual resale price maintenance and for a dilution of the role
and power of the Monopolies and Restrictive Practices Commission
(which from then on, until the introduction of merger policy in the
UK in the mid-1960s, was called simply Monopolies Commission
and dealt only with monopolies) as major concessions to business
interests. He acknowledges, however, that the structure of the Act
also reflected a willingness not to be dogmatic about restrictive prac-
tices and the recognition of the potential merits of some restrictive
arrangements.
Mercer (1995) provides another interesting, although perhaps more
controversial, account of the origins of the 1956 legislation, empha-
sizing the diversity of interests between different sections of the busi-
ness community, in particular between large, efficient, internationally
oriented firms and smaller, less efficient ones. She argues that the
spread of economic liberalism in the 1950s and the growing discon-
tent of the public and the press with restrictive practices were not the
only factors behind the passing of the 1956 Act, and that the lack of
opposition by larger firms to government plans for the introduc-
tion of cartel legislation was also a key factor. Thus Mercer points
out that some key concessions to industry, such as those mentioned
above, were more likely to benefit larger firms than smaller ones. She
also argues, however, that larger firms did not expect that the Act
would eventually lead to the abolition of a wide range of restrictive
practices. Her argument should not, therefore, be taken to imply that
many or most larger firms actually benefited from the introduction of
cartel legislation.
The Act was based on the principle that restrictive agreements were
in general adverse to the public interest unless shown otherwise. A
key feature was the adoption of a judiciary procedure for assessing
26 Chapter 2
By the end of 1959 some 2,240 agreements had been registered, about
half of which were of nationwide application. Most of them (approx-
Cartel Policy and Competition in British Industry 27
4. The actual number of agreements for any given industry has little significance, for
some industries registered only one national agreement covering all or most industry
products, while others registered many agreements, each corresponding to a narrowly
defined product or to a specific region of the country.
28 Chapter 2
One could add that the respectability of price-fixing was well es-
tablished among manufacturers and distributors during the inter-
war years and up to 1956, and, as indicated by various reports of
the Commission during the period 1948–1956, many businessmen
were bewildered by the argument that price-fixing could be thought
of as being against the public interest (see Hunter 1966, pp. 135–
136).
That is not to say, of course, that registration was universal: there
is evidence that a nonnegligible number of agreements were not
registered. Some of these continued secretly, while others were
abandoned. For instance, some of the agreements that had been con-
demned by the Monopolies and Restrictive Practices Commission in
its investigations of particular industries were never registered.
Since these had attracted a great deal of publicity, they were probably
abandoned shortly after the passing of the 1956 Act rather than
being secretly continued. Swann et al. (1974, pp. 153–154) mention
that out of sixty industries with restrictive agreements listed in the
Commission’s report on collective discrimination, eight did not
register their agreements. This is a nonnegligible percentage,
but it has to be borne in mind that in some of these industries the
agreements had comprised only the collective enforcement of resale
price maintenance or collective exclusive dealing, without any fixing
of common prices or market shares. Many of these would normally
not be registered, because the 1956 Act contained an outright pro-
hibition of the collective enforcement of resale price maintenance.5
Once this is taken into account, the incidence of nonregistration of
registrable agreements appears to have been rather low.
6. For example, one agreement was allowed to continue partly because it was said to
result in cost savings for buyers because there was no need to ‘‘shop around.’’ Another
was thought to promote product standardization. And still another was allowed to
stand because the Court accepted the argument that the restrictions helped keep prices
below the competitive level by reducing uncertainty, and hence the return on capital
required by firms in the industry.
7. Registration continued during the 1960s and the 1970s. These were mostly agree-
ments of minor significance or agreements that came into operation after the 1956 Act
had come into force.
30 Chapter 2
dustry, and the debate largely focused on the issue of the impact of
the removal of the minimum price scheme on the structure of the
industry, and hence on employment.
The association convinced the Court that the removal of the price
restrictions would be detrimental to employment in specific areas of
the country. At the same time, the Court accepted the Registrar’s
view that the agreement had resulted in prices higher than those
that would have prevailed under competition, and had allowed the
survival of inefficient producers and the maintenance of excess
capacity in the industry, all of which were against the public inter-
est. The Court struck down the agreement on the grounds that its
beneficial effect on employment was outweighed by the detriments
to the public, including the fact that it might have delayed what
seemed to be the necessary restructuring of a declining industry.
It is generally thought that the decision in the Yarn Spinners’ case
was an important factor in inducing a large number of industries to
voluntarily abandon their restrictive practices.
Another important Court case concerned a series of bilateral
agreements between steel firms and a marketing company that
these firms had formed to sell basic slag, a by-product of steel manu-
facture used as a fertilizer. Each of the steel producers appointed
one director to the board of the marketing company and had an inde-
pendent agreement with the company about the terms on which it
supplied it with basic slag, including the obligation to sell the whole
of its output of basic slag to the company. These bilateral agreements
were all identical, and although there was no explicit arrangement
between the steel firms, the Court thought that there was an implied
arrangement between them, which was registrable under the 1956
Act and against the public interest.
The importance of this case lies not in the economic significance of
the product in question, but in the fact that it settled the question
of what constitutes an implied arrangement in UK cartel law. The
Court pointed out that an arrangement between A and B exists if ‘‘(1)
A makes a representation as to his future conduct with the expec-
tation and intention that such conduct on his part will operate as an
inducement to B to act in a particular way, (2) such representation is
communicated to B, who has knowledge that A so expected and in-
tended, and (3) such representation or A’s conduct in fulfilment of
it operates as an inducement, whether among other inducements or
not, to B to act in that particular way.’’
Cartel Policy and Competition in British Industry 31
Two other important cases that came before the Court in the mid-
1960s involved arrangements to exchange information on prices and
price changes. Such arrangements were not registrable under the
1956 Act, unless it could be shown that they were being effectively
used as a way to continue a formally abandoned explicit collusive
scheme. In the two cases brought before the Court, there had con-
sistently been information exchanges regarding prices and price
changes in advance of their coming into effect, and the prices of dif-
ferent firms had always been identical. The Court concluded that
these information agreements amounted to the same effect as ex-
plicit price-fixing, and found against the parties to the agreements.
According to Swann et al. (1974), these Court decisions were in-
strumental in inducing several more industries with information
agreements to cancel them in the mid-1960s.
In fact, the extensive use of information agreements was causing
considerable concern among economists and policy makers in the
mid-1960s. Early studies of the impact of the 1956 Act showed that in
many industries explicit price-fixing agreements had been replaced
by informal arrangements to exchange information on prices, price
changes, quantities, and conditions of sale, and that in many cases
these had serious adverse effects on competition. Thus one of the
amendments to the 1956 legislation introduced by the 1968 Restric-
tive Trade Practices Act was a provision for making certain types of
information agreements registrable. It was left to the Board of Trade
to decide which types of information exchanges should be targeted,
however, and only agreements on the exchange of prices and condi-
tions of sale were subsequently called up for registration. By that
time most had already been abandoned.
A second deficiency of the 1956 Act related to the illegal non-
registration of agreements. The Act gave the Registrar of Restrictive
Trading Agreements only moderate powers of discovery. For in-
stance, it did not allow unexpected raids and seizing of company
documents. In addition, there were no significant penalties for non-
registration. In fact, Hunter (1966, pp. 87–88) has argued that the
main deterrent against nonregistration was probably the fear of ad-
verse publicity.
On the other hand, the 1956 Act did provide for penalties in cases
of secret collusion when the firms had been parties to an agreement
32 Chapter 2
that the Registrar had referred to the Court and the Court had struck
down, irrespective of whether the parties had actually defended the
agreement in Court. In some such cases considerable fines were
imposed. This provision was relevant not only for agreements that
had been voluntarily abandoned after having been formally referred
to the Court and condemned, but also for agreements that had been
abandoned before being referred to the Court and condemned. In this
latter case the referral was a precautionary step taken by the Regis-
trar in order to reduce the likelihood of any future anticompetitive
conduct by the firms involved. It is reasonable to assume that these
categories contained the majority of the most serious cases of collu-
sion. Still, a large number of abandoned agreements were not referred
to the Court, so the provision of the 1956 Act regarding penalties
would be irrelevant for firms that had been parties to these (as it
would be for firms that had not been parties to any registered
agreement). The 1968 Act sought to reduce the incentives for non-
registration by allowing injured parties to sue participants to un-
registered agreements for damages. This provision was rarely used,
and it is difficult to know the extent to which it may have deterred
nonregistration.
An Overall Assessment
8. This is also the view taken in the 1979 comprehensive government review of cartel
policy in the UK (SSPCP 1979). Symeonidis (1998) describes subsequent developments
in cartel legislation in the UK, including the legislative attack on restrictive practices in
the commercial services industries in the 1970s, and also provides a general assessment
of the law prior to the 1999 significant restructuring of UK competition policy as a
whole.
9. See Political and Economic Planning (1957) and Kuipers (1950) for the details on the
legal status of the restrictive agreements. The main point is that associations acting in
restraint of trade could not enforce in court agreements between members about con-
ditions on which they transacted business, or agreements to pay penalties. A member
who broke such agreements could not be sued.
34 Chapter 2
10. The only known case of an industry with restrictions on media advertising was
the linoleum industry, a homogeneous-good industry with very low advertising
expenditure.
11. Perhaps the only known case of clear and extensive cooperation in R&D was the
case of the permanent magnet industry, a small industry with modest R&D intensity
and little economic significance.
36 Chapter 2
The lack of excessive profits does not mean that the agreements were
not effective. The effectiveness of an agreement depended on two
factors: the extent to which the parties conformed to the agreement
or accepted the recommendations,14 and the extent of competition
from outside firms, domestic or foreign. Evidence on these issues
comes from various sources, including the agreements registered
under the 1956 Act; the Monopolies and Restrictive Practices
12. Cartelization in many British industries in the 1950s may have been associated
with high costs and low productivity rather than high profits. See Broadberry and
Crafts (1996) for an econometric analysis of the evolution of labor productivity across
British industries over the period 1954–1963 (which includes the first few years of the
implementation of the 1956 Act) that provides support for this view.
13. I will return to this issue in chapter 7, where I will present strong econometric
evidence consistent with the view that entry was, on the whole, not restricted in car-
telized industries.
14. While some of the agreements provided for sanctions such as a fine or expulsion
from the association in case of nonconformity, most agreements had no such provi-
sions and were therefore entirely voluntary. The provisions regarding fines were not
legally enforceable anyway.
38 Chapter 2
15. The limited extent of foreign competition across a range of UK industries in the
1950s is reflected in the low level of import penetration. A rough calculation at the
four-digit industry level suggests that, in more than 80% of manufacturing industries,
imports accounted for less than 10% of the total value of sales in the domestic market.
Even in industries with significant imports—for instance, 50% or more of total sales in
the domestic market—the high level of imports does not necessarily reflect competi-
tive pressure. In fact, in many industries with high import penetration, the imported
goods were largely complementary to domestic products within the relevant four-
digit industry (for example: preserved fruits and vegetables, oils and fats, nonferrous
metals, timber). In other cases the industries concerned were subject to quantitative
controls or other forms of regulation (thus imports in bacon, iron ore, and other prod-
ucts were very high, but they were a residual source of supply).
Of course, this should not be taken to imply that all UK manufacturing industries
were relatively free from foreign competition. Sectors facing a relatively low degree of
protection included most food and drink industries, some chemicals, basic metals,
clothing and footwear, wood products, publishing, leather and most textile semi-
manufactures, and building materials. Note, however, that in several of these indus-
tries, foreign competition was limited due to transport costs or international collusive
agreements. Sectors with a relatively high degree of protection included the engineer-
ing industries, instruments, vehicles, finished metal goods, some chemicals, paper and
paper products, furniture, pottery and glass, most finished textile goods, rubber prod-
ucts, and various other finished manufactures. See Political and Economic Planning
(1959), Morgan and Martin (1975), Kitchin (1976), and Milward and Brennan (1996) for
details on these issues.
40 Chapter 2
There have also been two more systematic studies of the impact of
the 1956 Act on competition. Heath (1961, 1963) conducted a ques-
tionnaire survey of the short-run consequences of the Act, using a
sample of 159 canceled price agreements. The survey was conducted
in 1960, only about a year after industries started formally abandon-
ing explicit price-fixing. In only about a third of the cases did the
respondents think that competition had intensified. This can be ex-
plained by short-term inertia or reluctance to abandon long-standing
practices. In fact, the results of the survey seem to strengthen the
view that the majority of industries had not anticipated the hard line
taken by the Restrictive Practices Court toward cartels, and were not
prepared for a more competitive environment in the short run. Heath
Cartel Policy and Competition in British Industry 41
With respect to the long run, Swann et al. found that in many indus-
tries with information agreements, price competition emerged
after these were abandoned or changed from prenotification into
postnotification arrangements in the mid-1960s, following the two
decisions of the Restrictive Practices Court on information agree-
ments discussed earlier in this chapter. As a result, price wars oc-
curred in a number of previously collusive industries in the second
half of the 1960s, and the final blow came with the provisions of
the 1968 Restrictive Trade Practices Act regarding information agree-
ments. Thus in most industries with information agreements, com-
petition emerged about a decade after the introduction of the 1956
legislation.
While the overall picture was one of a gradual strengthening of
competition in British manufacturing industries, Swann et al. also
noted that in the longer term there were instances of secret collusion
42 Chapter 2
16. The distinction between collusion and parallel pricing implied here is not just one
of form, but also one of effect. Thus effective collusion between a given number of firms
is assumed to always increase profit margins above ‘‘competitive’’ levels, while this is
not necessarily the case for parallel pricing.
Cartel Policy and Competition in British Industry 43
The 1956 Restrictive Trade Practices Act was quite effective in dras-
tically reducing the extent of cartelization in British manufacturing
between the 1950s and the 1970s. In particular, the large majority of
collusive industries registered their explicit restrictive agreements,
and then had to abandon them. The process was slow, and it took
more than a decade for the effects of the law to be fully realized.
Admittedly, there is some uncertainty about the impact of the legis-
lation on competition in each particular industry. Some agreements
may have not been effective at the time they were formally canceled,
and others may have been replaced, even in the longer term, by
secret or tacit arrangements. More generally, the ‘‘degree of collu-
sion’’ can vary, depending on the type of restrictions, the extent of
outside competition, the balance of interests within the cartel, and so
on—and so can the ‘‘intensity of competition’’ in the absence of
any explicit collusion. Is it then legitimate to reduce this continuum
of states to just two competition regimes and interpret the increase
in price competition following the cancellation of agreements as a
change in regime?
A cautious approach would be to say that for an industry that
abandoned an explicit restrictive agreement, there is a high proba-
bility that there was, at some point after the cancellation of the
agreement, a change of competition regime induced by the legisla-
tion, while this probability is zero for an industry not affected by
the Act. The empirical analysis carried out in this book, based on a
comparison of the evolution of the two groups of industries, would
still be valid under this interpretation. To the extent that some agree-
ments were ineffective in the 1950s and others were replaced by
17. Another finding of Swann et al. was that prices or price-cost margins in some of
the industries where competition broke out were restored or partly restored in the long
run. This does not imply, however, a weakening of competition. As described in sev-
eral of the case studies contained in Swann et al. (1973), this often occurred either after
the number of firms in the industry had decreased, in which case it is perfectly con-
sistent with competition remaining effective (see chapter 7 of this book), or because
prices had fallen to very low levels during a price war or due to a temporary fall in
demand.
44 Chapter 2
undetected secret or tacit collusion in the 1960s and the 1970s, the
analysis would, if anything, only somewhat understate the mag-
nitude of the effect of price competition on the endogenous vari-
ables of interest, namely, market structure, nonprice conduct, and
profitability.
I think, however, that there is no reason to emphasize this more
cautious interpretation. The survey and case-study evidence re-
viewed in this chapter suggests that the large majority of industries
that dropped their explicit collusive agreements as a result of the
1956 legislation did experience, sooner or later, an intensification of
price competition. Furthermore, the effect was in many cases sig-
nificant. Hence it is, on the whole, valid to think of this evolution
as a change of competition regime induced by an exogenous insti-
tutional change. Consequently, a comparison between those indus-
tries affected by the 1956 Act and those not affected should provide
an accurate assessment of the effects of competition.
3 The Competition Data
3.1 Introduction
industries that were investigated before 1956 and did not register
their agreements (or registered modified versions of them) must have
abandoned them (or abandoned the unregistered restrictions) soon
after the introduction of the legislation. It is indeed unlikely that in
these much-publicized cases explicit collusion would have secretly
continued. The situation with industries investigated after 1956 is
slightly more complicated. These were usually investigated because
of the presence of a dominant firm or a dominant group of firms. In
the majority of these cases, either there is no mention of any restric-
tive arrangements before or after 1956 or there is a reference to some
previously registered agreements. In a few cases, on the other hand,
an industry is reported as having had an agreement in the 1950s
which either was abandoned, and therefore never registered under
the 1956 Act, or was modified before registration. Finally, there are a
few cases where the Monopolies Commission uncovered an existing
secret collusive agreement during its investigation; information is
then provided on the origins of the arrangement.
For the purposes of my classification, all restrictive agreements
mentioned in the Commission’s reports as having been effective prior
to 1956 (or after that date) were treated in exactly the same way as
registered agreements—or, more precisely, as registered agreements
defended in the Court, since there was no question in those cases of
the restrictions not being effective.
Other reliable sources on the state of competition across manu-
facturing industries in the 1950s include the 1955 Monopolies and
Restrictive Practices Commission report on collective discrimination
(MRPC 1955), and several industrial case studies undertaken in the
mid-1950s as part of an investigation into the structure of British in-
dustry and contained in Burn (1958). Various other publications with
occasional references to competition in particular industries in the
1950s and the 1960s are mentioned in appendix A. There were only a
few agreements reported in these sources and not in any others, and
they were treated in the same way as registered agreements for the
purposes of my classification.
Finally, two very important but not fully reliable sources are the
Political and Economic Planning study of industrial trade associa-
tions, undertaken just before the 1956 Act was passed, and the Board
of Trade annual reports on the operation of the 1948 Monopolies and
Restrictive Practices Act, covering the years 1949 to 1956 (Board of
Trade 1949–1952, 1953–1956). These sources provide information on
The Competition Data 49
1. The reverse is also true: several industries that registered agreements do not appear
as collusive in the Political and Economic Planning survey or the Board of Trade an-
nual reports. On the whole, these two sources probably underestimate rather than
overestimate the extent of cartelization in British industry in the 1950s.
50 Chapter 3
One possible objection is that the data on British cartels relate to ex-
plicit collusion, but not to tacit collusion. Could it then be the case
that some of the industries classified as noncollusive actually prac-
ticed tacit collusion? One cannot, of course, rule out the possibility
that firms in an industry colluded tacitly in the absence of any ex-
plicit arrangement. However, it is difficult to understand why col-
luding firms would not want to enter into an explicit arrangement,
given that explicit collusion was legal and widespread in the 1950s.
Furthermore, the 1956 Act required the registration of informal and
even ‘‘implied’’ understandings as well as formal agreements, and
this seems to cover cases of tacit collusion. So tacitly colluding firms
The Competition Data 51
The above discussion suggests that the assumption that the existence
of an explicit price-fixing agreement is a good overall indicator of
collusive conduct is not an unreasonable one in the present con-
text. There is, however, one final difficulty, which is probably more
serious, for it relates to the issue of nonregistration. Again, there is a
weaker and a stronger version of the argument. The weaker version
is that nonregistration of agreements, if widespread, would lead to
serious measurement error. The stronger version is that certain types
of industries may have had a stronger incentive to avoid registration
than others. Then the failure to take unknown cases of collusion into
account could lead to sample selection bias.
There are two possible reasons for nonregistration: firms may sus-
pend an agreement, so that there is nothing to register, or they may
switch to secret or tacit collusion. Take, first, the former case. A rea-
sonable conjecture in this case is that very weak agreements would
be more likely to be dropped immediately than stronger ones. Even
if that were true, we would not be losing much by failing to identify
such cases, since we are primarily interested in identifying effective
agreements, not ineffective ones. But it is not even clear why the
decision to immediately cancel an agreement rather than register it
should have occurred in certain types of industries more than in
The Competition Data 53
2. It was in fact much larger in chemicals and food and drink, and only slightly larger
in textiles. However, the textiles sector is mentioned by Lydall as one with a high
proportion of firms acknowledging the existence of price leadership.
The Competition Data 55
manufacturing industries during the 1930s and the early 1940s. Given,
however, that most of the British cartels of the 1950s had been active
at the time of the Board of Trade surveys (see Swann et al. 1974), a
comparison of the information contained in the surveys with my
own data on collusion in the 1950s would certainly be interesting.
The two surveys differ somewhat in scope: the one on international
cartels was meant to be comprehensive, while the one on internal
cartels was conceived as the first part of a wider investigation, and
was therefore somewhat less comprehensive, although the industries
or sectors chosen spanned the whole spectrum of manufacturing
industries, covering capital-good industries as well as consumer-
good industries.
Of the two surveys, the one on internal cartels is probably the
more useful for my present purposes. The number of industries
reported in this survey as being subject to price-fixing and related
restrictions is equal to almost a third of the total number of indus-
tries classified as collusive in my data set (see table 3.1). In other
words, the coverage of the Board of Trade survey was quite large.
In spite of this large coverage, a careful reading of the report re-
vealed virtually no cases of industries reported as being subject to
restrictive agreements in the late 1930s or early 1940s which were not
mentioned as being collusive in any of my primary data sources
for this book. Furthermore, nearly all of these industries registered
agreements under the 1956 Act.3
An Overall Assessment
3. The Board of Trade survey of international cartels is less relevant for my present
purposes because in many cases only one British firm—usually a dominant firm in the
relevant UK industry—participated in the international agreement. Moreover, inter-
national cartels often broke down during or after the Second World War. Nevertheless,
the report also mentions several agreements in which more than one British firm was a
party. Most of these are also reported in my data sources for the 1950s, although a few
are not. It is impossible to know whether these agreements were still effective in 1956.
Even if they were, this would change my classification of industries in only three cases.
56 Chapter 3
Types of Restrictions
5. This is in sharp contrast with agreements that contained maximum or fixed dis-
counts to distributors or users in addition to the maintenance of individually set resale
prices. These agreements were often among manufacturers only. As mentioned above,
I have classified such schemes as significant restrictions on competition.
6. Admittedly, Yamey’s view has not been without its critics, partly because of evidence
that manufacturers in many industries opposed the abolition of resale price mainte-
nance at the time of the introduction of the 1964 Resale Prices Act. Still, this opposition
was attributed by most of these critics to the presumed efficiency gains from resale
price maintenance rather than to any potential effect of this practice on competition
between manufacturers (see Mercer 1998).
60 Chapter 3
8. The view expressed here is not inconsistent with the view taken by the Monopolies
and Restrictive Practices Commission in its report on electrical machinery (MRPC
1957). Although most sections of the electrical machinery industry had explicit price-
fixing agreements in the 1950s, a few had only agreements to report inquiries received
from prospective purchasers. The Commission concluded that this practice might be
expected to operate against the public interest because of the danger that under-
standings about prices or allocation of work would follow. Thus the Commission es-
sentially recognized the uncertain effect of such arrangements. In the context of the
heavily cartelized electrical machinery industry, however, the Commission was prob-
ably justified in stressing the potential for collusion. In other cases, the assessment
would have been less straightforward.
9. Note that I have not specifically discussed collective exclusive dealing and other
forms of collective discrimination, such as aggregated rebates. The reason is that these
were typically used as ancillary restrictions to strengthen a pricing agreement. How-
ever, in very few industries collective exclusive dealing was practiced even though the
manufacturers did not collude on prices or trade discounts. These cases present
something of a paradox: If collective exclusive dealing was being used as an entry
barrier, why didn’t these firms also attempt to restrict price competition among them-
selves? Since some of these cases are discussed extensively in reports of the Monopolies
and Restrictive Practices Commission, there can be no question of insufficient infor-
mation about these industries. A possible explanation is that these were instances of
bilateral collective exclusive dealing, the primary purpose of which was to protect the
interests of distributors rather than those of manufacturers. Fortunately, the industries
concerned are very few, and either they are not included in the statistical data sources
used for this book or they have been classified as ambiguous anyway.
62 Chapter 3
11. This also implies that the classification of marginal cases makes little difference to
the results. There were, in fact, very few marginal cases (i.e., cases where the fraction
of industry sales revenue subject to restrictions was either very close to 10% or 50%, or
was higher than one of these cutoff points in one year and lower in another year).
These cases were classified on the basis of two criteria. The first was the average or
typical value of this fraction over all relevant years. The second was an assessment of
which industry products were mainly driving the industry-level values of the endog-
enous variables of interest in each particular case. For instance, in the case of concen-
tration, I compared the average sales revenue of firms producing regulated products
with the average sales revenue of firms producing nonregulated products, the as-
sumption being that the market shares of the larger firms within an industry will
largely drive the concentration ratio of the industry as a whole.
64 Chapter 3
12. The potential selection bias due to exclusion of ambiguous industries may be more
of a concern for the results on the determinants of collusion presented in the final sec-
tion of the present chapter. However, as pointed out below, these results were found to
be robust to alternative classifications of the industries according to their competitive
status in the 1950s.
13. Except when (1) the fraction of the industry covered by the continuing agreement
was less than 50% or (2) the fraction of the industry covered by the continuing agree-
ment was more than 50% but another fraction of the same industry, between 10% and
50%, was affected by the Act. Such cases were classified as ambiguous. I should also
emphasize that the decisions by the Restrictive Practices Court to allow certain agree-
ments to continue do not appear to have been determined by any specific character-
istics of the industries involved (see the discussion in chapter 2). Nor was the evolution
of these industries similar during the 1960s. Take, for instance, the two collusive
industries included in my sample for chapter 4 and assigned to the group that was not
affected by the 1956 legislation: cement and glazed tiles. In the cement industry the
five-firm concentration ratio increased from 0.86 in 1958 to only 0.89 in 1968, while the
price-cost margin increased from 0.25 to 0.29 during the same period. On the other hand,
in the glazed tile industry the five-firm concentration ratio increased from 0.65 in 1958
to 0.94 in 1968, and the price-cost margin fell from 0.19 to 0.18 during the same period.
66 Chapter 3
Export Agreements
A small complication that has arisen with the above method of clas-
sifying the industries relates to the operation of export agreements.
As mentioned in chapter 2, export agreements were not registrable
under the 1956 Act. Thus information on whether any given collu-
sive industry had a restrictive agreement for the export market in
addition to the one for the home market is often not available. Fur-
thermore, little is known about whether and when any particular
export agreement was abandoned. Now if exports in a collusive in-
dustry were never subject to restrictions, or if they were subject to
restrictions but remained so after 1958, then the impact of the 1956
legislation on competition in the industry would be smaller than
what we would measure it to be under the assumption of no differ-
ence between the home market and the export market. Only in the
case where the restrictions originally applied to both markets and
were subsequently dropped in both markets would we be measuring
things correctly.
It could be argued that this is indeed what happened in many,
probably most, cases. It is clear from the available information that
a large number of agreements covered exports as well as domestic
sales. Also, once collusion had broken down in the domestic market,
it may have been difficult for firms to maintain it in the export market,
especially in the context of intensifying international competition
of the 1960s and the 1970s. To this argument one could add that
the whole issue is of no importance for industries with insignificant
exports.
Still, the issue should not be dismissed so easily. In a few cases,
we know that the export market was not subject to collusion, and
we can take this into account when calculating the fraction of in-
The Competition Data 67
Broadberry and Crafts (1996, 2000), who have carried out a separate
survey of restrictive agreements in British industry in the 1950s, have
68 Chapter 3
used somewhat different criteria than the ones that I use here to clas-
sify industries as collusive, competitive, or ambiguous in the 1950s.
As a result, their categorization of industries is different from mine
in several cases, although in almost all of these the difference is by
one class (i.e., collusive versus ambiguous, or ambiguous versus
competitive). Broadberry and Crafts use the three-digit level of aggre-
gation, while I typically use more disaggregated industry categories.
A side effect of this is that the divergence between our respective
classifications is smaller than it would have been if I had used
the three-digit level of aggregation as well. More specifically, there
are three main differences between their classification criteria and
mine.
First, Broadberry and Crafts rely heavily on registered agreements,
and make relatively little use of other sources on collusion in British
industry. For instance, they do not mention the Monopolies and Re-
strictive Practices Commission reports or the Board of Trade annual
reports. They do report some information from the Political and Eco-
nomic Planning survey, however, although it is not clear how they
are using this information when classifying the industries. A com-
parison of their lists of industries with mine suggests that this factor
accounts for part, although not most, of the divergence between our
respective classifications.
Second, their two principal criteria for assessing the significance of
registered agreements are (1) the existence of a detailed price list
with nationwide application, and (2) whether the agreement was re-
ferred to the Restrictive Practices Court by the Registrar or at least
specifically mentioned in the Registrar’s reports as being a pricing
agreement. These criteria are somewhat different from the ones that I
use to classify the significance of particular types of restrictions (see
above). For instance, I do not regard price-fixing agreements that
contain a detailed price list as being necessarily more significant or
more effective than ones which do not. Clearly, the choice of criteria
for assessing the significance of particular agreements is to some ex-
tent subjective. As it turns out, however, the difference between the
Broadberry and Crafts approach and my approach in this respect
does not result in any major discrepancies between our respective
classifications of industries, partly because most of the ambiguous
cases end up in the intermediate category both in their categorization
and in mine. Nevertheless, this factor adds to the divergence be-
tween our respective classifications.
The Competition Data 69
Table 3.1
Competition and collusion across manufacturing sectors in the 1950s
No. of industries classified as
Collusive Ambiguous Competitive
Mining and quarrying 4 0 0
Food, drink, and tobacco 8 8 14
Chemicals and allied products 1 4 8
Basic metal manufacture 12 1 2
Mechanical engineering 6 10 8
Instrument engineering 0 3 2
Electrical engineering 12 2 5
Shipbuilding and vehicles 0 4 2
Other metal products 9 4 2
Textiles 8 9 9
Leather, leather goods, and fur 0 0 4
Clothing and footwear 0 1 11
Bricks, pottery, glass, cement, etc. 8 4 1
Timber, furniture, etc. 1 1 4
Paper, printing, and publishing 5 2 3
Other manufacturing industries 2 3 5
Total 76 56 80
70 Chapter 3
14. In constructing this sample, I have excluded four industries with significant gov-
ernment participation or intervention in the 1950s (sugar, ordnance and small arms,
aircraft, locomotives), two industries that hardly existed prior to the mid-1950s (syn-
thetic rubber, tufted carpets), and one where imports were very much higher than
domestic production and were subject to regulation (iron ore).
15. In comparing the two tables, recall that some collusive industries experienced no
change in regime, and for others the evolution was ambiguous because of agreements
with uncertain effect on competition that were made in the 1960s. Another reason for
the discrepancies in the classification of industries across the two tables is that for table
3.1 the 10% and 50% cutoff points are applied to the period 1951–1958, while for table 3.2
they are applied to the period after 1958. The correlations reported in section 3.4 below
are based on table 3.1, but this of course is not the case for the results of later chapters.
The Competition Data 71
Table 3.2
The evolution of competition across manufacturing sectors after 1958
No. of industries with
Change in No change in
competition competition
regime between regime between
1958 and 1975 1958 and 1975
Mining and quarrying 1 0
Food, drink, and tobacco 8 15
Chemicals and allied products 1 8
Basic metal manufacture 12 2
Mechanical engineering 5 8
Instrument engineering 0 2
Electrical engineering 11 5
Shipbuilding and vehicles 1 2
Other metal products 8 3
Textiles 7 9
Leather, leather goods, and fur 0 4
Clothing and footwear 0 11
Bricks, pottery, glass, cement, etc. 5 3
Timber, furniture, etc. 1 4
Paper, printing, and publishing 3 3
Other manufacturing industries 2 5
Total 65 84
The data on the British cartels of the 1950s provide a very rare op-
portunity to examine the impact of several structural industry char-
72 Chapter 3
Theoretical Background
Firms that collude have generally two problems to solve. They must
coordinate on a price or a set of prices. And they must ensure that
collusion is enforceable, that is, that none of the parties has an in-
centive to cheat. Thus a common distinction in the literature on the
factors facilitating or hindering collusion is between coordination
and enforcement of collusive agreements.
Product homogeneity, high industry concentration, the absence
of significant cost or size differences between firms, institutional
features such as the existence of an industry trade association, and
sociocultural characteristics such as ‘‘homogeneity of values’’ and
social ties among senior managers are often cited as factors facilitat-
ing coordination on a collusive price or set of prices (see, e.g., Scherer
and Ross 1990). In addition, coordination must be easier under a
permissive climate of competition policy or when collusive agree-
ments are not illegal. While these ideas are intuitively appealing,
they have proved difficult to model within a formal game-theoretic
framework.
The Competition Data 73
16. See Fershtman and Pakes (2000) for a recent theoretical approach to collusion that
allows for entry and exit.
74 Chapter 3
17. There is also an empirical literature on the related but different issue of cartel
duration (for instance, Jacquemin et al., 1981; Suslow, 1991; Dick, 1996b). Probably the
most robust result from this line of research is the negative effect of demand uncer-
tainty on cartel stability.
18. Going beyond this binary classification—for instance, by classifying industries with
respect to the ‘‘degree of collusion,’’—is not possible, given the information available.
The Competition Data 75
19. Since there were various possible reasons for classifying an industry as ambiguous,
it did not seem appropriate to regard these industries as intermediate with respect to
the state of competition and include them in the sample, except as a mere robustness
check. As discussed in Symeonidis (1999b), this has no effect on the results.
20. These are the only available concentration data for the 1950s that correspond to the
industry categories I will be using to analyze the effects of the 1956 legislation on market
structure in later chapters.
21. In some cases, a four-digit industry may comprise two or more submarkets which
are relatively independent on the demand side. I identified eight such cases in the
sample. Three were in the collusive group, and five in the competitive group, so it is
unlikely that any bias is caused from this factor.
76 Chapter 3
Statistical Evidence
22. The idea is that while the actual level of the advertising-sales ratio and the R&D-
sales ratio may depend on many variables, including the intensity of price competi-
tion, it is generally exogenous industry characteristics that will determine whether
these ratios are above or below 1% (or 2%). Thus, for an industry below the 1% cutoff
point, advertising/R&D is not an important strategic variable: in such an industry,
advertising is not very effective in raising consumers’ willingness to pay or there is
little scope for technological innovation from within the industry. On the other hand,
in an industry above the 1% cutoff point (and even more in an industry above the 2%
cutoff point), advertising/R&D ‘‘works.’’ Of course, whether such an industry has an
advertising-sales or R&D-sales ratio of 5% or 10% may be largely determined by
endogenous factors. However, my binary variables AD1 and RD1 are not very sensi-
tive to endogenous factors that affect advertising intensity and R&D intensity.
I should also point out that a comparison of advertising-sales ratios and R&D-sales
ratios across various years revealed very few instances where an industry had moved
from below 1% to above 1% or vice versa; and in most cases this was due to an exog-
enous institutional change, the introduction of TV advertising in the UK in the 1950s.
78 Chapter 3
Table 3.3
Descriptive statistics for collusive and noncollusive industries, 1958
Collusive Noncollusive
ðn ¼ 71Þ ðn ¼ 80Þ
C5 Mean 0.61 0.52
Standard deviation 0.20 0.28
Minimum 0.19 0.08
Maximum 0.98 0.99
ln K=L Mean 1.23 0.64
Standard deviation 0.73 0.86
Minimum 0.99 1.18
Maximum 3.10 2.38
ln K=N Mean 0.01 1.09
Standard deviation 1.15 1.33
Minimum 2.98 3.60
Maximum 2.17 3.42
No. of industries with AD1 ¼ 0 62 49
No. of industries with AD1 ¼ 1 9 31
No. of industries with RD1 ¼ 0 59 60
No. of industries with RD1 ¼ 1 12 20
Note: n indicates the number of industries.
23. A Wilcoxon rank-sum test of the hypothesis that the distribution of C5 is the same
in the two groups also rejects the null at the 5% level but not at the 1% level.
The Competition Data 79
Table 3.4
Correlation coefficients between COLL and other variables, 1958
Correlation coefficient (significance level)
COLL and C5 0.19 (0.02)
COLL and ln K=L 0.34 (0.00)
COLL and ln K=N 0.40 (0.00)
COLL and AD1 0.29 (0.0002)
COLL and RD1 0.10 (0.23)
Table 3.5
Advertising, R&D, and collusive pricing, late 1950s
No. of No. of
collusive noncollusive
industries industries
intensity and setup costs in this group. I will return to this issue
below.
Note also the strong and highly significant negative correlation
between COLL and AD1, suggesting that collusion is less likely in
advertising-intensive industries than in industries with low adver-
tising intensity, as well as the negative relationship between COLL
and RD1, which is not, however, statistically significant at the 5%
or even the 10% level. A more detailed breakdown of industries ac-
cording to their competition regime, advertising intensity, and R&D
intensity is provided in table 3.5. Note that none of the industries in
the sample with ADS higher than 5% is collusive.
To check whether there is any (linear) association between con-
centration and collusion once we control for capital intensity, I have
also computed the partial correlation coefficient between COLL and
C5, keeping ln K=L constant. It is equal to 0.007, with a significance
level of 0.93. On the other hand, the partial correlation coefficient
between COLL and ln K=L, keeping C5 constant, is equal to 0.294,
80 Chapter 3
A possible reason for the absence of any clear link between concen-
tration and the incidence of collusion in the present context is that
the coordination and monitoring of collusion must have been greatly
facilitated by the fact that the agreements were not illegal and were
often operated by members of trade associations. In fact, the admin-
istration of collusive agreements was one of the primary functions of
British industrial trade associations in the 1950s, and often their main
function (see Political and Economic Planning 1957). On the other
hand, a nonlinear relationship between concentration and collusion
could result from the fact that high levels of concentration are often
associated with significant firm asymmetries or the presence of dom-
inant firms, and these factors may hinder collusion.25
24. Thus, although there is no reason to doubt that the growth of tariffs in the 1920s
and the 1930s had greatly facilitated the operation of price agreements, as argued by
Swann et al. (1974, p. 197), it is not clear whether tariff protection was an important
determinant of the incidence of collusion across industries in the 1950s.
25. One argument which was used by the authors of the 1946 Board of Trade report on
British cartels to explain the absence of restrictive agreements in several industries
where a single firm controlled a very large part of the market is that in these industries
there may be little scope for collusion, since prices are effectively decided by the
dominant firm. However, such cases are not very numerous at the four-digit industry
level—although they may be more numerous at a lower level of aggregation. More-
over, some industries with dominant firms are not included in the present data set,
because their concentration ratio is not reported in the official statistics to avoid dis-
closure of information about specific firms.
82 Chapter 3
26. Differences in product quality between firms will also be small in industries with
process rather product R&D. This may also weaken any link between R&D intensity
and collusion, although it needs to be said that several industries where process R&D
is important, such as basic chemicals, are not included in the sample used here.
The Competition Data 83
This chapter has prepared the way for the analysis of the effects
of the 1956 legislation in the rest of the book. In the first part, I have
described the available data on cartels and competition in British in-
dustry from the 1950s to the mid-1970s. I have argued, on the basis of
84 Chapter 3
4.1 Introduction
1. It is equally difficult to interpret the evidence of a positive link between price com-
petition and concentration in US exogenous sunk cost industries reported by Robinson
and Chiang (1996), since their measure of competition was based partly on demand
characteristics, such as the degree of product homogeneity, and partly on very specific
factors that were assumed to affect firms’ conduct, such as the infrequency of orders.
Price Competition and Concentration 89
split into four groups: a group with agreements upheld by the Court,
another with agreements struck down by the Court, a third group
comprising industries which voluntarily abandoned their agreements,
and a control group of industries which, according to the authors,
had not been subject to collusion prior to the introduction of the 1956
Act. Three different measures of merger activity were used: the ratio
of expenditure on acquisitions to total assets, the proportion of firms
in an industry taking over other firms, and the proportion of firms in
an industry being acquired. The authors found no statistically signif-
icant difference in the ratio of expenditure on acquisitions to total
assets among the four groups, and they also found that the control
group had the highest proportion of both acquirers and acquired
firms. They concluded that there was no evidence of any effect of the
1956 Act on merger activity during the period under study. How-
ever, their sample may have been too small to be representative.
Also, the criteria used to classify industries to the four groups were
not always clear.2
In summary, the empirical evidence from previous studies of the
effect of competition on concentration in the UK, and also in the US,
has been inconclusive or mixed. This chapter undertakes a systematic
and extensive reexamination of the UK evidence, focusing on exoge-
nous sunk cost industries. Later chapters in the book will examine
the evidence for advertising-intensive and R&D-intensive industries.
Overall, I extend the previous studies of the effects of the 1956 Act on
market structure in several ways. First, I relate my empirical analysis
to a theoretical framework, two key elements of which are the dis-
tinction between classes of industries and the focus on jointly deter-
mined endogenous variables. Second, I look at the impact of the Act
over a longer time period, from the mid-1950s to the mid-1970s, and
for the whole of manufacturing industry. Third, I use more compre-
hensive data on competition; in particular, I use information from
several sources to identify unregistered agreements and cases of col-
lusion in the 1960s. Fourth, I use explicit criteria to classify industries
according to their competitive status. Fifth, I pay much attention to
the issue of potential selection or endogeneity bias when interpreting
the evidence. And, finally, I perform regression analysis to control
2. For instance, some of the industries included in the control group had in fact been
subject to restrictive agreements in the 1950s, and one of the industries included in the
group whose agreements were upheld by the Court dropped the agreement shortly
after the Court’s decision.
Price Competition and Concentration 91
3. Advertising and R&D are probably not the only endogenous sunk costs incurred by
firms. For instance, the cost of developing a network of contacts or relationships may
also be thought of as an endogenous sunk cost. The reason for ignoring it in the pres-
ent book is twofold. First, it is probably less important (or provides less scope for
endogenous investments by firms) than advertising or R&D, at least in the manu-
facturing sector. Second, it is difficult, if not impossible, to measure in a systematic
way across industries.
92 Chapter 4
4. Alternatively, one could assume that firms always achieve the highest level of col-
lusion that is sustainable, given a number of parameters taken as exogenous at the
‘‘short-run’’ competition stage. Under this interpretation, a rise in t might correspond
to a lower critical discount factor in an infinitely repeated game.
Price Competition and Concentration 93
At stage 1 of the game, firms enter as long as the profit they expect to
make at stage 2 covers the cost of entry. Hence the equilibrium level
of N is given, in the symmetric case, by the largest integer satisfying
Extensions
entry on price may often be modest (see Bresnahan and Reiss 1991, Geroski 1991,
Sutton 1997a).
To see this, define S ¼ Npq, where q is output per firm, and notice that firm profit
can be written as p ¼ ð p cÞq ¼ S=N cq, where c is marginal cost. Then suppose that
N rises, which implies that pq falls, given S. Now if p is nonincreasing in N, then p
must fall when N rises, whatever happens to q: if q falls, p ¼ ð p cÞq also falls, while
if q rises, p ¼ S=N cq again falls. A similar argument can be made with respect to
changes in S. As for t, notice that a rise in t (i.e., a movement away from joint profit
maximization and toward the one-shot Nash equilibrium) causes p to fall and q to rise
so that pq remains constant, given S and N; hence p ¼ S=N cq falls.
Price Competition and Concentration 95
7. Note that h has a direct equivalent in standard models of horizontal product dif-
ferentiation: it would be the inverse of the degree of product substitutability in a non-
address model of product differentiation, such as Dixit and Stiglitz (1977), or the
consumers’ unit transport cost (or intensity of preferences) in an address model of
product differentiation without choice of location, such as Salop (1979). Both these
models can be seen as special cases of the general framework presented here, since h
is taken in these models as exogenous and pi is increasing in h. On the other hand,
there is a sharp distinction between horizontal product differentiation and advertising
in the present theory—unlike much of the literature following the structure-conduct-
performance approach, in which advertising was very often used as a proxy for the
degree of product differentiation. Advertising is modeled in chapter 5 as a source of
vertical (as opposed to horizontal) product differentiation.
8. See Norman and Thisse (1996) and d’Aspermont and Motta (2000) for an analysis of
the positive effect of price competition on concentration using address models of hor-
izontal product differentiation. Section 4.3 below illustrates the positive competition-
concentration link in the context of a nonaddress model of product differentiation.
96 Chapter 4
9. The adjustment need not be slow if the cost of the capital investment is spread
across many years through interest payments or depreciation allowances. Mergers
intended to restore the market power of firms following a rise in the intensity of com-
petition could also speed up the adjustment of concentration to its new long-run equi-
librium value. Another factor that could speed up the adjustment is the fact that more
efficient firms may acquire less efficient ones soon after competition intensifies to avoid
pronounced price-cutting behavior. In turn, less efficient firms, which feel vulnerable
once competition intensifies, may prefer to be acquired immediately (i.e., before com-
petition seriously affects their balance sheet position).
98 Chapter 4
(1991, 1998) has proposed a ‘‘bounds’’ approach for testing this class
of predictions.10 However, it is difficult or inappropriate to apply a
bounds approach in the present empirical context, for reasons that
will be discussed later in this chapter. The alternative is to carry out a
regression analysis of the competition effect on market structure, on
the implicit assumption that the theoretical results obtained for
minimum concentration apply to actual concentration as well. In
other words, it will be assumed that the mechanism identified
for the benchmark case of symmetric single-plant, single-product
firms dominates any other effects that may arise in a more complex
setting.11
Clearly, the validity of such a claim must be tested against the em-
pirical evidence. However, one remark may be made here. The key
issue is whether the symmetry assumption is appropriate in light
of the existence of considerable asymmetries between firms in any
given industry. These asymmetries often reflect efficiency differ-
ences. Now one might expect that once price competition intensifies,
high-cost firms (which were previously making zero or almost zero
net profit) will be more easily driven out or taken over, while low-
cost firms (which were previously making positive net profit) will be
more likely to stay in the industry and expand through growth or
mergers. At the new long-run equilibrium, the marginal firm will
again be making zero or almost zero net profit. Hence, in the ab-
sence of any initial negative correlation between firm size and
efficiency—which seems quite implausible for the typical industry
—concentration should rise. In particular, the expansion of low-cost
firms will only reinforce the effect of falling firm numbers on con-
centration. This story is perfectly consistent with the basic intuition
from the symmetric case. In fact, the model presented in this section
captures the basic mechanism driving the evolution of market struc-
10. The bounds approach, as described in Sutton’s work, has mostly focused on ana-
lyzing the properties of lower bounds to concentration. Recent theoretical work by
Nocke (2000) has extended this approach to the study of upper bounds to concentration
in different classes of industries.
11. A more complex theoretical framework for the analysis of market structure would
combine deterministic factors deriving from game-theoretic models of entry (such as
the one described in this chapter) with stochastic patterns of firm growth that may ac-
count for much of the asymmetry observed in real markets (such as those analysed in
Hart and Prais 1956, Simon and Bonini 1958). Sutton (1998) presents an integrated
theory of the determinants of market structure that combines the two approaches.
Price Competition and Concentration 99
12. According to a stronger version of the efficiency differences approach, the number
of firms need not fall even though concentration should rise following an intensifi-
cation of price competition. In other words, changes in concentration would be mostly
driven by changes in the distribution of market shares rather than by changes in firm
numbers. This hypothesis is discussed in detail in chapter 7. However, it is rejected by
the empirical evidence reported in that chapter.
100 Chapter 4
X XX
U¼ ða1 x i a2 xi2 Þ a2 s x i xj þ M: ð4:4Þ
i i j<i
X
pi ¼ a1 2a2 x i a2 s xj ð4:5Þ
j0i
in the region of quantity spaces where prices are positive. Let N de-
note the number of varieties offered; this is also the number of firms
that have entered the industry. Inverting the system of N equations,
we derive the consumer’s demand function for variety i:
P
ða1 pi Þ½2 þ sðN 2Þ s ða1 pj Þ
j0i
xi ¼ ð4:6Þ
a2 ð2 sÞ½2 þ sðN 1Þ
in the region of prices where quantities are positive. It can be easily
14. This demand system was introduced by Bowley (1924). A different type of qua-
dratic utility function, which includes the number of firms as a parameter, was pro-
posed by Shapley and Shubik (1969) and was used in slightly different form in Shubik
(1980). See Martin (2002) for a detailed comparison of the two models. Sutton (1997b,
1998) provides a comprehensive analysis of a quantity-setting oligopoly using the
Bowley demand system. This seems a natural choice in the present context, where the
number of firms is endogenous, and hence its inclusion as a parameter in the utility
function seems inappropriate.
15. See Shaked and Sutton (1990) for a discussion of how such a utility function can
also be derived by aggregating the preferences of heterogeneous groups of consumers.
102 Chapter 4
16. The parameter l was first suggested by Edgeworth (1881), who called it the ‘‘co-
efficient of effective sympathy’’ and applied it to the study of individual behavior. It
has been used in oligopoly models by Cyert and deGroot (1973), who named it ‘‘coef-
ficient of cooperation,’’ Kuenne (1980), and Shubik (1980), among others. Note that l is
free from some of the theoretical problems encountered in other approaches to mod-
eling the intensity of short-run competition by way of a reduced-form parameter, such
as the conjectural variations approach.
What also justifies the use of l as a reduced-form competition measure is its prop-
erties in the final-stage subgame: it can be checked that, for given N, the equilibrium
price, price-cost margin, and profit in the second-stage subgame increase and the
equilibrium quantity falls as l rises (the degree of collusion increases). None of the
other exogenous variables that affect gross profit, namely, s or c, have properties sim-
ilar to those of l in the second-stage subgame. Of particular interest in this respect
are the properties of s, since this has often been used as a measure of the intensity of
competition. It can be checked that a fall in s (i.e., an increase in the degree of product
differentiation) increases both the equilibrium price and the equilibrium quantity of the
second-stage subgame.
Price Competition and Concentration 103
a1 c
xðNÞ ¼ :
a2 ½4 þ sð1 þ lÞðN 1Þ
The profit of each firm is then given by
Sða1 cÞ 2 ½2 þ slðN 1Þ
pðNÞ ¼ : ð4:7Þ
a2 ½4 þ sð1 þ lÞðN 1Þ 2
It can be easily checked that pðNÞ is increasing in S, decreasing in s,
17. Within this more general formulation, industry A would be defined as being more
collusive than industry B if l A ðNÞ > l B ðNÞ, EN.
104 Chapter 4
pðN Þ ¼ f ; ð4:10Þ
where N is the long-run equilibrium number of firms. Given (4.8)
and (4.9), it is then easy to see from the total differential of equation
(4.10) that, for any l A ½0; 1Þ, dN =dl > 0 (i.e., the equilibrium number
of firms increases as the degree of collusion rises). For l ¼ 1 (i.e.,
at the perfectly collusive point), pðNÞ is maximized and, given (4.8)
and the exogenously fixed f , so is N . Thus more price competition
raises concentration in an exogenous sunk cost industry. It is also
easy to check that dN =dS > 0 and dN =ds < 0, that is, the number of
firms increases with market size and falls as the products become
less differentiated.
It is straightforward to check that the assumption of quantity-
setting firms is innocuous in this model, for similar results are de-
rived for the case of price-setting firms. Moreover, the results do
not depend on the assumption of single-product firms either, because
similar results are obtained when we allow each firm to produce M
varieties, at least for the case where M is exogenously fixed and is
the same across firms.
Concentration
18. Recall that in the system of Census industry definitions, the industry categories
become progressively narrower as digits are added to the industry code. Thus each
three-digit industry includes a number of four-digit industries, each of these includes a
number of five-digit industries, and so on.
106 Chapter 4
19. The industry categories in the official concentration statistics are defined in accor-
dance to a number of criteria, the most important of which are (1) that they do not
have very low sales, and (2) that they are reasonably homogeneous.
I will also use a few official concentration ratios for 1968 reported in Hart and Clarke
(1980) but not published in the official statistics, and a few concentration ratios for
1975 taken from Statistics of Product Concentration of UK Manufactures for 1975, 1976 and
1977, Business Monitor PO 1006 (London: H.M.S.O., 1980). I will not use concentration
ratios for 1976–1977, even though such data are available from the above source and/
or from a list of unpublished concentration ratios for 1976–1977 compiled by officials
at the Office of Fair Trading and kindly made available to me by David Elliott. My
reason for not using the 1976–1977 data in addition to the 1975 data is to avoid creat-
ing a panel with a very asymmetric intertemporal structure. I could, of course, use the
1977 (or 1976) figures instead of the 1975 figures. This would barely affect the results,
as can be seen by comparing the results reported here with those reported in Symeo-
nidis (2000a). I have chosen to use 1975 instead of 1977 (or 1976) here for two reasons:
first, because this ensures the maximum possible comparability with other data sets
used in this book, and, second, because in a few cases 1976–1977 concentration data
comparable to those for previous years are not available.
20. For most of the period examined in this book, the plant-level data underlying the
Census of Production are not available, because the individual returns to the Census of
Production have, unfortunately, been destroyed.
Price Competition and Concentration 107
21. Whenever, on the other hand, explicit use is made of the Cournot model of oli-
gopoly, the Herfindahl index may be a more appropriate concentration measure (see
Cowling and Waterson 1976). This is not the case here. In any event, the Herfindahl
index and the concentration ratio are very strongly correlated. Sawyer (1971), who
estimated approximate four-firm concentration ratios and Herfindahl indices for three-
digit British manufacturing industries in 1963, found a correlation coefficient of 0.94
between the two measures.
108 Chapter 4
22. Moreover, imports were, on the whole, not significant in the 1950s, but increased
during the 1960s and the 1970s. Hence differences across industries in the evolution of
the import-adjusted concentration ratio would be influenced by differences across
industries in the extent of import penetration.
Price Competition and Concentration 109
23. In particular, note that summary table 5 of the 1963 Census of Production contains
data for 1958 and 1963, while Statistics of Product Concentration of UK Manufactures for
1963, 1968 and 1975 contains data for 1963, 1968, and 1975. Now noncomparability was
usually the result of a small change in industry definition from one table to the other
rather than from one year to the next within a single table. Hence concentration (and
sales revenue) figures were available for 1963 under both definitions. In such cases,
and provided that the change in industry definition was minor, I applied a simple
proportional adjustment to some of the published figures to ensure comparability over
time. To make sure, in such cases, that the difference in the 1963 figures across tables
was due to a change in industry definition rather than to the revision of a figure, I
checked the individual Census industry reports.
24. Sometimes the noncomparability was due to a change in industry definition from
one year to the next within the same table. Census industry reports were then exam-
ined to assess the magnitude of the change. Usually it was minor (less than 10%);
concentration ratios were then left unchanged—although sales figures were adjusted
on the basis of information from the Census industry reports. Whenever the change
was large, the noncomparable figures were discarded.
110 Chapter 4
Market Size
25. Nearly all of these were cases where the total number of firms in the industry
(excluding those with fewer than twenty-five employees) was slightly larger than five.
In order to avoid the disclosure of information on particular firms, the concentration
statistics then report an n-firm concentration ratio equal to 1 (where n > 5 is the total
number of firms).
26. For example, with a constant-elasticity demand function the total number of profit
opportunities in the market is simply proportional to the value of sales, for any given
degree of collusion; hence, the first proxy may be preferable. With a linear demand
function, however, the second proxy may be preferable, especially when the changes
in relative prices are mainly driven by changes in relative marginal costs rather than
by shifts in demand.
Price Competition and Concentration 111
tion. Using each of these proxies in turn and comparing the results
has one further advantage in the present case: it strengthens the case
for treating sales revenue as an exogenous variable in these regres-
sions (this point will be further discussed in the next section).
Data on the value of manufacturers’ sales at current net producer
prices at the four-digit level of aggregation were obtained mostly
from the same publications as the concentration ratios, and in some
cases from the industry reports of the 1958, 1963 and 1968 Censuses
of Production and the 1975 issues of Business Monitor, PQ series:
Quarterly Statistics. The figures relate to sales by all plants (for 1968
and 1975) or firms (for 1958 and 1963) employing at least twenty-five
persons. The exclusion of very small plants is probably a minor issue,
since these often produce ancillary rather than core industry prod-
ucts. Moreover, a search through the individual industry reports of
the Census of Production did not reveal any serious discontinuities
caused by the switch from excluding very small firms to excluding
very small plants for any industry in my sample.27
A series of general producer price indices was obtained from the
Annual Abstract of Statistics. Producer price indices for a number of
four-digit industries, covering all or part of the period under study,
have been published in the Annual Abstract of Statistics, in various
issues of the Board of Trade Journal and Trade and Industry, in the
Annual Bulletin of Construction Statistics, and in Business Monitor, PQ
series: Quarterly Statistics. For many industries, however, data are not
available or are not available for all years, so price indices at a higher
level of aggregation or of closely related industries have sometimes
had to be used as proxies. An alternative set of price indices can be
constructed on the basis of data on the volume of output often
reported, along with data on sales revenue, in the individual industry
reports of the 1958, 1963, and 1968 Censuses of Production and in
Business Monitors.28
Both these sets of price indices have shortcomings in addition to
the fact that they are often incomplete. Published price indices are
calculated largely from information supplied by firms and trade
27. Such discontinuities were present in only four industries: granite, limestone, sand
and gravel, and ready-mixed concrete. In these industries, firms that employed more
than twenty-five persons and operated one or more small plants each accounted for a
significant fraction of total sales. None of the four industries is included in any of the
samples used in this book, however, because they were all classified as ambiguous
with respect to their state of competition after 1958.
28. See Nicholson and Gupta (1960) for an early application of this method.
112 Chapter 4
29. Whenever both these problems were present, the choice usually was between a
published index for a three-digit or a two-digit or a closely related four-digit industry,
Price Competition and Concentration 113
The problems associated with the use of the published sales data
were largely similar to those mentioned above for the concentration
data. Thus data were often not available for all years. On the other
hand, noncomparabilities caused by small changes in industry defi-
nitions were usually dealt with by adjusting some of the published
figures.
Competition
on the one hand, and an imperfect output-based price index, on the other. I usually
chose the latter approach in the case of sectors where industries are largely hetero-
geneous with respect to costs (e.g., the food and drink sector), and the former for
sectors where a greater degree of homogeneity with respect to costs is present while
output figures are often problematic because of the high degree of product differ-
entiation (e.g., the engineering sector). Sometimes I computed both indices and took an
average.
114 Chapter 4
Setup Cost
There are no data for setup costs, so two different proxies were con-
structed. The first of these was constructed by defining a measure of
minimum efficient scale relative to industry size and multiplying it
by the total value of capital stock in the industry (see Sutton 1991).
I used the simplest possible measure of minimum efficient scale,
namely, the size of the average plant. Divided by industry size, this
becomes equal to the inverse of the number of plants in the industry.
Hence the first proxy used for setup cost is the capital stock of the
average plant. A measure of minimum efficient scale based on the
30. An alternative approach to modeling the competition effect was used in Symeoni-
dis (2000a, 2000b). It involved constructing, on the basis of certain plausible assump-
tions about the timing of the impact of the legislation, an industry-year-specific dummy
variable that takes the value 1 for ‘‘collusive’’ and the value 0 for ‘‘competitive’’
industry-year pairs. This alternative approach imposes more structure on the data and
does not allow for an analysis of the short-run impact of the 1956 Act. On the other
hand, it can take into account some information about dates when particular agree-
ments were abandoned. In any case, the two approaches yield very similar results re-
garding the long-run effect of the Act on concentration (see Symeonidis 2000a).
Price Competition and Concentration 115
31. In fact, the inclusion of time effects in the empirical specification relaxes even more
the assumptions on the parameters yi , gi , and di , since they are also allowed to change
over time, albeit in a uniform way across industries.
116 Chapter 4
32. Part of the cost of plant and machinery may also be recoverable on exit. As pointed
out above, however, differences across industries in the degree to which setup costs
are sunk will be largely captured in panel regressions by the industry effects.
33. A small complication with the 1975 figures has arisen from the fact that the
O’Mahony and Oulton (1990) capital stock estimates for 1979 are not adjusted to con-
trol for the effect of the apparently considerable amount of premature scrapping of
capital assets after 1973. However, the authors provide an estimate for the extent of
premature scrapping for all manufacturing industry after 1973. I therefore used this
estimate to adjust my 1975 capital stock figures, on the assumption that the extent of
premature scrapping was constant for every year between 1973 and 1979 and similar
across industries. Since the relevant period for applying this adjustment in my data set
is very short, the resulting figures for 1975 are very similar to the unadjusted ones.
Price Competition and Concentration 117
There are two main problems with the raw data, both of which are
acknowledged by the authors.
The first problem is that the investment data before 1948 are less
reliable and are available only at a very high level of aggregation,
namely, for eleven industry groups covering the whole of manu-
facturing. The use of these data to construct capital stock figures for
1954 and later years introduces a potentially significant measure-
ment error, especially for 1954 and 1958. On the other hand, for 1963
and later years the influence of the pre-1948 investment on current
capital stock is negligible.
The second limitation relates to the way investment is defined in
the Census of Production. The Census figures are for acquisitions
minus disposals of capital assets, but the definition of disposals
refers only to the selling of assets; it does not include the scrapping
of assets either by firms that exit the industry or by remaining firms.
Of course, O’Mahony and Oulton assume finite asset lives and allow
for depreciation, so ultimately any scrapped assets are eliminated
from the capital stock estimates. But this happens with a significant
lag, and there may be measurement error induced by the fact that
certain industries experience a higher rate of exit or more rapid
technological change (leading to the scrapping of assets) than others
during a certain period.
The potential effect of these measurement errors on the results of
interest in this book is not clear a priori. Fortunately, it may not be
very significant in practice, as is suggested by the similarity of results
using the benchmark estimates of capital stock to results using some-
what more refined estimates of capital stock whenever such more re-
fined estimates could be constructed (see chapter 7, section 7.4, for a
more detailed discussion of this important point).
Data on the number of plants (or Census ‘‘establishments’’) at the
three-digit level of aggregation were taken from the Census of Pro-
duction (various years). There were a number of practical problems
in constructing a series of comparable figures for plant numbers, and
a brief discussion of these problems is now in order.
First, the total number of plants in an industry is greatly affected
by the number of very small plants. These typically account for a
very small fraction of industry capital stock and often produce an-
cillary rather than core industry products. An additional complication
in the present context is that the reported number of plants employing
118 Chapter 4
34. Whenever the 1963 industry definition had not been significantly modified in 1968,
I constructed 1958 figures according to the following formula: Estimated N(1958) ¼
½N1 (1958)=N1 (1963) N(1963), where N1 (1958) is the number of plants with at least
twenty-five employees in 1958, as published in summary table 4 of the 1958 Census;
N1 (1963) is the number of plants with at least twenty-five employees in 1963, as pub-
lished in summary table 4 of the 1963 Census; and N(1963) is the number of plants
with at least twenty-five employees in 1963, as published in summary table 1 of the
1968 Census (the reference table for the series as a whole).
Whenever the 1963 industry definition had been significantly modified in 1968, it was
usually still possible to construct comparable figures for 1958 with the help of more
disaggregated data for 1958 and 1963 published in the individual industry reports of the
1963 Census. These were data on the number of plants operated by firms employing at
least twenty-five employees, and were used on the assumption that the proportional
change in this number over the period 1958–1963 would be equal to the proportional
change in the number of plants employing at least twenty-five persons. I therefore
computed Estimated N(1958) ¼ ½N2 (1958)=N2 (1963) N(1963), where N2 (1958) and
N2 (1963) are the figures from the 1963 Census industry reports.
Price Competition and Concentration 119
35. In the large majority of industries, the proportional change in the number of
employees over the period 1968–1970 was small, while the proportional change in the
reported number of plants was larger and always negative, which is consistent with
what was said above about the noncomparability of the 1968 and 1970 figures for plant
numbers. In only a few cases was the reverse true, and I then used the reported figures
for plant numbers, on the assumption that the noncomparability problem was not
serious in these particular industries.
Note that the use of the proportional change in the number of employees as a proxy
for the actual proportional change in the number of plants captures changes in the
number of plants caused by the shutdown of plants or the settingup of new ones, since
these also affect employment. However, measurement error is induced by the fact that
employment in existing plants may also change. But the period 1968–1970 is such a
small part of the overall period examined in this book that this should not make much
difference to the estimated overall change in K/N across industries.
36. More specifically, I started by computing Estimated N(1970) ¼ ½L( 1970)=L( 1968)
N(1968), where L(1970) and L(1968) are numbers employed in plants with at least
twenty-five employees in 1970 and 1968, respectively, and N(1968) is the number of
plants with at least twenty-five employees in 1968, as published in summary table 1 of
the 1968 Census. Then I computed Estimated N( 1975) ¼ ½N1 (1975)=N1 (1970) Estimated
N(1970), where N1 (1975) is the number of plants with at least eleven employees in 1975,
as published in the 1975 Census industry reports, and N1 (1970) is the number of plants
with at least eleven employees in 1970, as published in the 1970 Census industry
reports.
120 Chapter 4
R&D-Sales Ratios
R&D expenditure data for the UK are available for various years
since 1964 at a level of aggregation between the two-digit and the
three-digit (for about thirty to forty ‘‘subsectors’’). They have been
published in Research and Development Expenditure, Studies in Official
Statistics no. 21 (London: H.M.S.O., 1973); Research and Development:
37. In subsequent chapters I also use data on capital stock, the number of plants, and
employment for 1954 and 1973. The capital stock data are from O’Mahony and Oulton
(1990). Employment figures were taken from summary table 1 of the 1958 Census and
from summary table 2 of the 1973 volume of Business Monitor, PA series: Report on the
Census of Production. Data on plant numbers were taken from the industry reports of
the 1958 and the 1973 Censuses. Corrections were made to ensure comparability over
time.
Price Competition and Concentration 121
38. It is not clear to what extent royalties constitute an endogenous sunk cost. In ad-
dition, there are serious data limitations regarding royalties paid and received by UK
firms at the industry level. On the whole, royalty payments were probably roughly
equal to royalty receipts for UK manufacturing as a whole during the period under
study, although payments were greater than receipts in engineering and the reverse
was true for chemicals.
122 Chapter 4
Advertising-Sales Ratios
39. In those cases where the level of aggregation in the Annual Line of Business reports
was higher than the UK four-digit industry level, the assumption was usually made
that all four-digit industries within a particular Line of Business category had the same
R&D intensity.
Price Competition and Concentration 123
Ltd., from 1963 until the early 1970s; and the MEAL Monthly Digest of
Advertising Expenditure, published since 1968 by Media Expenditure
Analysis, Ltd. These data are reported both for individual brands
and for industries, at a relatively low level of aggregation, but they
are often not available for low-advertising industries. Also, they re-
late to the UK market rather than to UK firms. Finally, data on ag-
gregate advertising expenditure by type of advertising have been
published by the Advertising Association in Advertising Expenditure
1960 (London: Advertising Association, 1962) and, subsequently, in
Advertising Quarterly.
A detailed comparison of these various data sources suggested (1)
that the Census figures probably include some expenditures that
represent sales promotion rather than media advertising, and (2) that
the Legion/MEAL figures are, on the one hand, downward-biased
due to incomplete coverage or noncoverage of certain media and, on
the other hand, upward-biased because they are estimated by pricing
the measured amount of advertising according to published rates,
that is, without taking discounts into account (this mainly affects the
estimates for TV advertising).
For my present purposes, advertising expenditure is defined as
media advertising only, since other selling expenses generally affect
marginal cost and are incurred at the price competition stage. The fol-
lowing procedure was therefore adopted for deriving advertising-sales
ratios at the four-digit level for industries covered by Legion/MEAL.
A minimum ADS was derived using Census sales data, approxi-
mately adjusted for net imports according to the Annual Statement
of the Trade of the United Kingdom, and Legion/MEAL advertising
data, adjusted downward to account for discounts from published
rates for media advertising. Changes to this minimum ADS across
time were checked. Moreover, a maximum ADS was estimated using
Census sales and advertising expenditure data. Since these data are
available for industry groups and only for 1963 and 1968, estimates
at the four-digit level were derived by using the Legion/MEAL fig-
ures as a guide for the relative advertising intensities of four-digit
industries within each industry group and by assuming that changes
over time were similar to those of the minimum ADS.
Now as pointed out above, the difference between the two esti-
mates of ADS is due to incomplete coverage or noncoverage of cer-
tain media by Legion/MEAL, and probably also to the inclusion of
some sales promotion expenditures in the Census figures. Although
124 Chapter 4
methods is possible only when the least squares residuals are posi-
tively skewed; otherwise the maximum likelihood estimates are
actually the same as the least squares estimates (see Waldman 1982).
As it turns out, the residuals from the least squares dummy variable
models estimated in this section are negatively skewed. This is due
to the fact that some industries experienced large increases in con-
centration during the period examined, so the observations for the
early years have large negative residuals. There may, however, be a
more fundamental problem with the estimation of a lower bound in
the present case, which is due to the panel structure of the data. In
particular, it may not be appropriate to control for industry effects
when estimating a bound. On the other hand, the failure to control
for industry effects would essentially reduce the data set to a pooled
time-series cross section, and it would then be very difficult to iden-
tify a competition effect on concentration because of the prevalence
of industry effects.
The above discussion suggests that the use of standard least squares
regressions is the most appropriate approach in the present case. Note
that an implicit assumption made under this approach is that the the-
oretical predictions for minimum concentration also apply to actual
concentration (i.e., that the mechanism driving the theoretical results
in the case of symmetric single-plant, single-product firms dominates
any other effects that may arise in a more complex setting).
The Sample
Table 4.1
The basic sample of exogenous sunk cost industries
No. of industries with No. of industries without
change of competition change in competition
Data available for regime after 1958 regime after 1958
1958, 1963, 1968, 1975 27 34 (of which 2 collusive)
1958, 1963, 1968 9 3
1963, 1968, 1975 1 15
1958, 1963 0 14 (of which 8 collusive)
1963, 1968 0 4
Total 37 70
Notes: (1) The group of industries without change in competition regime after 1958
includes two industries (glazed tiles and cement) in which collusion continued
throughout the period and for which data are available for all four years.
(2) The group without change in competition regime also includes eight industries
that were affected by the nationalization of steel in 1967. All of these had agreements
that were abandoned after 1963. They have been classified in this group because only
1958–1963 data are used for these industries.
(3) One industry with a change in competition regime and data available for 1963–
1975 is included in the sample, since this industry abandoned its agreement after 1963.
Table 4.2
Concentration in 1958 and competition regime in exogenous sunk cost industries
Mean C5 in 1958
(st. deviation of C5)
Basic sample
Industries with change of competition regime after 1958 0.514 (0.206)
(n ¼ 36)
Industries without change of regime after 1958, including 0.496 (0.281)
8 industries where collusion was abandoned after 1963
(n ¼ 51)
More balanced sample
Industries with change of competition regime after 1958 0.514 (0.206)
(n ¼ 36)
Industries without change of competition regime after 0.435 (0.277)
1958 (n ¼ 37)
Notes: (1) The figures are based on industries with available data for 1958; n indicates
the number of industries.
(2) The group of industries without a change in competition regime includes two
industries where collusion continued throughout the 1960s and the 1970s.
(3) The group of industries without a change in regime in the basic sample (but not in
the more balanced sample) includes eight industries that were affected by the nation-
alization of steel in 1967. If these industries are excluded, the mean C5 in 1958 for the
group without a change in competition regime in the basic sample is 0.455 (and the
standard deviation is 0.283).
the results are substantially the same when this asymmetry between
the two groups is removed. Thus I have also constructed a second
sample, which includes only industries with a full set of observa-
tions for 1958–1968; in the large majority of cases, the 1975 obser-
vation is also available. This ‘‘more balanced sample’’ contains 73
industries, including 36 with a change in competition regime, and
280 observations.
Descriptive Statistics
Table 4.3
Average change in C5, exogenous sunk cost industries: 1958–1968 and 1958–1975
DC5 1958–1968 DC5 1958–1968 DC5 1958–1975
Industries with 0.129 0.125 0.164
CHANGE ¼ 1 (0.117) (0.123) (0.158)
n ¼ 36 n ¼ 27 n ¼ 27
Industries with 0.087 0.085 0.115
CHANGE ¼ 0 (0.090) (0.090) (0.121)
n ¼ 37 n ¼ 34 n ¼ 34
Notes: The figures in the first column are based on seventy-three industries with
available observations for both 1958 and 1968. The figures in the second and third
columns are based on sixty-one industries with available observations for 1958, 1968
and 1975. The figures in parentheses are standard deviations. n indicates the number
of industries.
The econometric model that I will use is a panel data model with
individual-specific effects. These should control for industry-specific
characteristics that are relative stable, and thus unlikely to change
significantly for any given industry within a ten-year or a twenty-
130 Chapter 4
year period. Three time dummies, for 1963, 1968, and 1975, are also
included among the regressors (the benchmark year is 1958). There
are good reasons for this. The average five-firm concentration ratio
across UK manufacturing industries, which had already been rising
throughout the 1950s, increased by eight percentage points—from
about 55.5% to 63.5%—between 1958 and 1968, then changed very
little between 1968 and the mid-1970s (Hart and Clarke 1980, Hart
1985). Now it was during the 1958–1968 period that the impact of
the 1956 Act on competition was mostly realized. If the rise in con-
centration was partly caused by factors other than the 1956 Act or
changes in setup costs and market size, then the estimated coef-
ficients would be seriously biased if the model were specified with-
out time dummies.
Several factors have been cited as having contributed to the rise in
concentration during the 1960s. They include changes in the tax sys-
tem in the mid-1960s that are thought to have encouraged mergers;
economies of scale in product development, distribution, and the
raising of finance; improvements in the system of transport; the
emergence of large retailers with significant bargaining power; and
the progressive opening of the British economy. It is very difficult to
measure these factors at the industry level, and no such attempt will
be made here. However, there is no reason to expect that they would
affect the group of industries with a change in competition regime
and the control group of industries in different ways. To the extent
that their effect has been more or less realized across all industries, it
should be largely captured by the time dummies.
Of all these factors, the one whose omission from my empirical speci-
fication is the most regrettable is the intensification of foreign compe-
tition caused by the gradual opening of the British economy during
the 1960s and the 1970s. Unfortunately, it is difficult to control for this
in a more satisfactory way. Ideally, one would need some measure of
the extent of foreign competition for each industry across time. Two
possible candidates are the import penetration ratio and the rate of
effective protection. However, there are serious problems, theoretical
and practical, with both of these measures. Estimates of effective
rates of protection are available at a high level of aggregation and for
only some years in my sample; also, they are often subject to mea-
Price Competition and Concentration 131
40. In contrast, the Dillon Round tariff cuts of 1961 were small and were partially off-
set by the import surcharge of 1964–1966.
132 Chapter 4
in the other six. For industry groups that I could classify as having
experienced no change in competition regime, the respective num-
bers were eight and ten.
In summary, then, it is not unreasonable to argue that the esti-
mated effect of the 1956 Act in this and in subsequent chapters is not
biased by the failure to control for foreign competition.
41. Recall that since the model includes industry-specific effects, one need not assume
that the three-digit industry K/L or K/N is an accurate measure of setup cost at the
four-digit industry level. All that is required is that the change in K/L or K/N is roughly
similar for all four-digit industries within any given three-digit industry, and is an
accurate measure of the change in setup cost.
134 Chapter 4
42. It could also be argued that if lagged concentration were indeed an important
explanatory variable in the model, its omission would show as significant serial
correlation in the residuals. But there is no such evidence in the models estimated
below.
Price Competition and Concentration 135
43. In fact, it is by no means clear that the breakdown of collusion will lead to a rise in
industry sales revenue when the deflator used is the general producer price index for
all manufacturing. Whether sales revenue increases or decreases will depend on the
elasticity of demand. Although a joint monopoly would operate on the elastic part of
the demand curve, its abolition may well move the industry to the inelastic part of the
demand curve, so sales revenue may rise or fall relative to its original level. Moreover,
collusion need not involve joint profit maximization, so it is not even clear that a col-
lusive industry will be operating on the elastic part of the demand curve.
136 Chapter 4
44. Arguments similar to those that I use below, and in subsequent chapters, to defend
the key identifying assumption in my analysis of the effects of the 1956 Act have also
been put forward in several other recent studies of natural experiments. See, for in-
stance, Chevalier (1995), Eissa (1995), and Eissa and Liebman (1996).
Price Competition and Concentration 137
ever, there are comparable figures for firm numbers in both 1954 and
1958 for a large sample of industries. As will be pointed out in chapter
7, where these data are analyzed in detail, there was almost no differ-
ence whatsoever in the evolution of firm numbers during the period
1954–1958 between the two groups of industries.45
Further evidence can be obtained from a comparison of three-firm
employment concentration ratios for three-digit industries in 1951
and 1958. The data are contained in Armstrong and Silberston (1965),
and are taken from Evely and Little (1960) and the 1958 Census of
Production. Because of changes in industry definitions and other
noncomparabilities between the two years, the sample is small. The
average change in the three-firm concentration ratio C3 during the
period 1951–1958 was 2.5 percentage points for fifteen exogenous
sunk cost industries with a change of competition regime after 1958,
and 2.0 percentage points for eleven exogenous sunk cost industries
without such a change. A test of independence of the respective
means of the change in C3 in the two groups could not reject the null
hypothesis of independence at the 5% level or, indeed, at the 10% or
20% level. Although this sample is very small and the three-digit
level of aggregation is far from ideal, there is clearly no evidence of
any differential underlying trend between the two groups.
Econometric Results
45. A four-year period is, of course, rather short. However, the period 1958–1963
is also rather short, and this does not prevent the difference between the two groups
(presumably caused by the 1956 Act) from emerging sharply during that period. See
chapter 7 for details.
138 Chapter 4
46. The reported results are all for models without serially correlated disturbances. I
also estimated a model with an AR1 error structure, and the results were very similar to
those reported here. The estimated coefficient of serial correlation in these regressions
was very low (and sometimes even negative): between 0.05 and –0.12 when using the
basic sample, and between 0.05 and –0.03 when using the more balanced sample.
Price Competition and Concentration 139
140
Regression results for concentration in exogenous sunk cost industries (fixed-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.078 — 0.082 — 0.30 — 0.31 —
(0.016) (0.017) (0.10) (0.10)
ln DS — 0.073 — 0.073 — 0.31 — 0.31
(0.015) (0.016) (0.09) (0.10)
ln K=N 0.089 0.100 — — 0.45 0.49 — —
(0.023) (0.023) (0.14) (0.13)
ln K=L — — 0.022 0.035 — — 0.19 0.23
(0.038) (0.038) (0.21) (0.20)
Y63 0.006 0.004 0.027 0.026 0.07 0.07 0.16 0.16
(0.012) (0.013) (0.014) (0.014) (0.08) (0.08) (0.08) (0.08)
Y68 0.038 0.035 0.081 0.078 0.21 0.21 0.39 0.39
(0.020) (0.020) (0.025) (0.025) (0.11) (0.11) (0.14) (0.13)
Y75 0.038 0.033 0.100 0.093 0.23 0.22 0.48 0.47
(0.027) (0.027) (0.038) (0.039) (0.16) (0.16) (0.21) (0.21)
CHANGE Y63 0.041 0.040 0.038 0.036 0.09 0.09 0.08 0.07
(0.020) (0.020) (0.021) (0.021) (0.12) (0.12) (0.12) (0.12)
CHANGE Y68 0.062 0.056 0.057 0.050 0.26 0.24 0.24 0.22
(0.021) (0.021) (0.021) (0.021) (0.12) (0.12) (0.12) (0.12)
Chapter 4
CHANGE Y75 0.073 0.062 0.067 0.056 0.32 0.27 0.30 0.25
(0.026) (0.026) (0.027) (0.027) (0.14) (0.14) (0.15) (0.15)
R2 0.52 0.51 0.50 0.48 0.45 0.46 0.44 0.44
141
Table 4.5
142
Regression results for concentration in exogenous sunk cost industries (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.085 — 0.076 — 0.38 — 0.33 —
(0.012) (0.013) (0.07) (0.07)
ln DS — 0.081 — 0.071 — 0.38 — 0.33
(0.012) (0.013) (0.07) (0.08)
ln K=N 0.123 0.124 — — 0.60 0.61 — —
(0.010) (0.010) (0.06) (0.06)
ln K=L — — 0.140 0.141 — — 0.72 0.72
(0.015) (0.015) (0.08) (0.08)
Y63 0.002 0.001 0.002 0.003 0.04 0.05 0.05 0.06
(0.012) (0.012) (0.013) (0.013) (0.07) (0.07) (0.07) (0.08)
Y68 0.020 0.024 0.019 0.022 0.14 0.16 0.12 0.14
(0.013) (0.013) (0.016) (0.016) (0.07) (0.07) (0.08) (0.08)
Y75 0.008 0.013 0.006 0.002 0.10 0.13 0.002 0.02
(0.016) (0.016) (0.020) (0.021) (0.09) (0.09) (0.11) (0.11)
CHANGE Y63 0.040 0.040 0.041 0.040 0.09 0.09 0.09 0.09
(0.018) (0.018) (0.018) (0.019) (0.10) (0.10) (0.10) (0.10)
CHANGE Y68 0.062 0.055 0.067 0.060 0.26 0.23 0.29 0.26
(0.018) (0.018) (0.020) (0.019) (0.11) (0.11) (0.11) (0.12)
Chapter 4
CHANGE Y75 0.076 0.064 0.085 0.074 0.34 0.28 0.39 0.34
(0.022) (0.022) (0.023) (0.023) (0.13) (0.13) (0.13) (0.13)
Constant 1.486 1.443 1.170 1.119 4.46 4.49 2.91 2.89
143
Table 4.6
144
Regression results for concentration in exogenous sunk cost industries: more balanced sample (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.084 — 0.076 — 0.37 — 0.33 —
(0.014) (0.014) (0.07) (0.08)
ln DS — 0.078 — 0.067 — 0.37 — 0.32
(0.014) (0.015) (0.08) (0.08)
ln K=N 0.122 0.126 — — 0.61 0.62 — —
(0.013) (0.013) (0.07) (0.07)
ln K=L — — 0.124 0.127 — — 0.68 0.68
(0.021) (0.021) (0.11) (0.11)
Y63 0.003 0.003 0.003 0.002 0.01 0.01 0.02 0.02
(0.014) (0.014) (0.016) (0.016) (0.08) (0.08) (0.09) (0.09)
Y68 0.024 0.027 0.033 0.035 0.15 0.16 0.16 0.17
(0.016) (0.016) (0.019) (0.020) (0.08) (0.08) (0.10) (0.10)
Y75 0.011 0.013 0.012 0.014 0.11 0.12 0.06 0.07
(0.019) (0.020) (0.025) (0.026) (0.10) (0.10) (0.13) (0.13)
CHANGE Y63 0.042 0.042 0.044 0.044 0.12 0.12 0.13 0.13
(0.020) (0.020) (0.020) (0.021) (0.11) (0.11) (0.11) (0.11)
CHANGE Y68 0.059 0.051 0.062 0.055 0.26 0.23 0.28 0.25
(0.020) (0.019) (0.021) (0.021) (0.11) (0.11) (0.12) (0.12)
Chapter 4
CHANGE Y75 0.074 0.062 0.080 0.069 0.34 0.28 0.37 0.32
(0.023) (0.024) (0.024) (0.025) (0.14) (0.14) (0.14) (0.14)
Constant 1.469 1.403 1.175 1.080 4.38 4.34 2.87 2.70
145
146 Chapter 4
47. Another problem with the behavioral interpretation is that it may not always be
valid. I will return to this issue in chapter 7, section 7.3.
This page intentionally left blank
5 Price Competition,
Advertising, and Market
Structure in Advertising-
Intensive Industries
5.1 Introduction
2. Schmalensee (1992) has argued that it is not clear that advertising should be seen as
a longer-term decision than price, because advertising depreciates relatively quickly
and price, too, can often have durable effects on demand. He also shows, however, that
key theoretical results derived from a three-stage game that treats advertising as a
sunk cost are robust to alternative approaches to modeling the interaction between
advertising and price competition, provided that advertising is treated as a fixed cost.
3. This is the approach taken in Symeonidis (2000b).
Price Competition, Advertising, and Market Structure 153
4. Note that since dpi /dAi > 1 at Ai ¼ 0, firms always find it profitable to advertise.
One could also imagine that the function pi is first convex and then concave in Ai .
Although a firm would never optimally choose a point on the convex part of the gross
profit function representing a positive amount of advertising, it might choose not to
advertise at all, if the net profit from doing so were larger than the net profit at the
point where dpi =dAi ¼ 1. Since the focus of this chapter is on advertising-intensive
industries, it seems natural to ignore cases that involve zero advertising. I also abstract
from the issue of potential threshold effects (in the sense that advertising below a cer-
tain level has little impact), since any part of the gross profit function below the
threshold would have little relevance for the present analysis.
5. This point is further elaborated in the next chapter, section 6.2, where the issue of
spillovers is discussed in more detail in the context of R&D-intensive industries. The
key insights from that discussion are relevant for advertising spillovers as well.
Price Competition, Advertising, and Market Structure 155
I begin with the first of three general results regarding the joint be-
havior of A and N (or the concentration ratio C ) following a change
in the intensity of price competition t. Note that analogous results can
be derived for the effect of a change in any of the other exogenous
variables in the model. Let DA and DN ðDC Þ denote the change
in A and N ðC Þ, respectively. The following proposition places
bounds on the space of outcomes by excluding certain outcomes as
not admissible within the theory.
We will also need to define the profit of firm i at the initial equilib-
rium values of t and N when firm i sets Ai ¼ A1 while all other firms
set Aj ¼ A0 :
p 0 ¼ pðAi ¼ A1 j S; N0 ; h; e; t0 ; Ai ¼ A0 Þ:
The proof is by contradiction. Assume that N1 b N0 and A1 b A0
following a rise in t (i.e., we also have t1 > t0 ). From A2 and the fact
that pi is decreasing in N and t, we obtain p1 < p 0 . Moreover, since
the function pi ðAi j S; N0 ; h; e; t0 ; Ai ¼ A0 Þ is concave in Ai and its
slope is equal to 1 at Ai ¼ A0 , its slope is smaller than 1 at all points
between A0 and A1 . Hence p 0 p0 < A1 A0 . Combining the two
inequalities, we obtain p1 p0 < A1 A0 . This is, however, impos-
sible, since net profit must be zero at equilibrium and therefore we
must have Dp ¼ DA . Hence it cannot be the case that DA b 0 and
DN b 0 following a rise in t.
To prove part (2) of the proposition, we only need to show that no
other outcomes can be excluded even for a much more structured
version of the model. The following set of ancillary assumptions is
intended to eliminate some of the indirect effects following a change
in t:
Assumption A4. dpi =dAi is nonincreasing in Aj , Ej 0 i.
Assume the opposite, that is, A0 A1 < p0 p 00 , given that equi-
librium concentration falls or does not change following a rise in t.
From assumption A3 and the assumption that pi is decreasing in
N , we obtain p1 a p 00 . Combining the two inequalities, we obtain
A0 A1 < p0 p1 , which is impossible, since net profit must be zero
at long-run equilibrium and hence we must have Dp ¼ DA . It fol-
lows that if concentration falls or does not change following a rise
in t, then it must be the case that A0 A1 b p0 p 00 . This completes
the proof. Note that the reverse is not true, that is, the condition
A0 A1 b p0 p 00 is not sufficient to rule out an increase in concen-
tration following a rise in t. (For instance, concentration may rise if
the fall in A reduces gross profit below p 00 .) r
6. A numerical example may help clarify this point. Consider an industry with
symmetric firms and unit-elastic demand, so that the industry sales revenue is fixed
and equal to S. Let ð p cÞ=p denote the price-cost margin and NA/S the industry
advertising-sales ratio. The free-entry condition implies ð p cÞ=p ¼ NA=S þ Nf =S,
where f is the exogenous sunk cost.
Suppose that initially ð p cÞ=p ¼ 6% and NA=S ¼ 1%. Now if the price-cost margin
falls below 5% as a direct result of tougher price competition, concentration must rise
because A cannot fall below zero. In other words, at any new equilibrium we must
162 Chapter 5
have NA=S b 0, so the free-entry condition can be restored only through a rise in the
price-cost margin and/or a fall in Nf/S, both of which require a fall in N. Note that
even with a smaller fall in the price-cost margin, a large proportional change in the
advertising-sales ratio is required if concentration is not to rise.
Things are different for high-advertising industries. Thus, for instance, if initially
ð p cÞ= p ¼ 6% and NA=S ¼ 4%, then a fall of the price-cost margin to 5% can be easily
offset by a fall in A, and concentration need not rise. In particular, if the price-cost
margin is homogeneous of degree zero in the Ai ’s, then N will remain unchanged pro-
vided that NA/S goes to 3%, which represents only a 25% decrease in the advertising-
sales ratio; and it will rise if A falls by more than 25%. If, on the other hand, a fall in A
causes a reduction in the price-cost margin, then A will need to fall by more than 25%
to keep N constant; but this will still be a modest proportional fall in A.
Price Competition, Advertising, and Market Structure 163
in the region of quantity spaces where prices are positive, and her
demand function for variety i is
P
ui ð1 pi Þ½2 þ sðN 2Þ s uj ð1 pj Þ
xi j0i
¼ ð5:5Þ
ui ð2 sÞ½2 þ sðN 1Þ
in the region of prices where quantities are positive, where N denotes
the number of firms that have entered the industry. It can be easily
seen that x i is decreasing in pi and increasing in pj , Ej 0 i. Also, it is
increasing in ui and decreasing in uj , Ej 0 i.7 Since there are S con-
sumers, firm i sells quantity Sx i . Let each firm have a constant mar-
ginal cost of production c, where c < 1.
As in chapter 4, I will use here a continuous reduced-form collu-
sion parameter to model different degrees of competition in the final
stage of the game. In particular, I assume that in the final stage each
firm maximizes the sum of its own (gross) profit and a fraction l of
the (gross) profit of each of the rival firms. The parameter l, 0 a
l a 1, is, then, an inverse measure of the intensity of short-run com-
petition, with l ¼ 0 representing the Cournot-Nash equilibrium and
l ¼ 1 corresponding to perfect collusion.
Although the use of l is a somewhat arbitrary way to model col-
lusion, it turns out to be very useful in providing insight about the
effects of price competition in advertising-intensive industries within
7. A nice feature of the present model—and one that distinguishes it from models of
‘‘pure vertical differentiation,’’ which do not have this property—is that it gives rise to
reduced-form profit functions where profit increases in own quality and decreases in
rival quality.
166 Chapter 5
8. The properties of l in the final-stage subgame are the same as in the version of the
model without quality indices described in chapter 4: for given N and for ui ¼ u, Ei, the
equilibrium price, price-cost margin, and profit in the second-stage subgame increase
and the equilibrium quantity falls as l rises. None of the other exogenous variables
that affect gross profit, namely, s or c, have properties similar to those of l.
9. No attempt is made here to explain what determines b and why it differs across
industries. Advertising effectiveness is simply taken as an industry-specific character-
istic which is not affected by firms’ strategies. This is consistent with the fact that the
industries where advertising ‘‘works’’ tend to be fairly similar across countries and
over time.
10. As pointed out in Sutton (1998), the economics of this model depends only on the
composite mapping from firms’ fixed costs to firms’ gross profits, rather than on the
separate mappings of fixed costs to qualities and from qualities to gross profits. Thus
the results would not change if we replaced u with u g , g > 0, in the utility function and
required that b > 2g.
Price Competition, Advertising, and Market Structure 167
and
E4 ¼ sð1 þ lÞðN 1Þ
are all nonnegative expressions. From (5.7), and using also (5.8), we
obtain
E3 ðE1 E2 Þ þ E1 ðE3 E4 Þ ¼ b½2 þ lsðN 1Þ½4 sð1 þ lÞ 2 ; ð5:9Þ
which defines the equilibrium number of firms N . Given N , (5.8)
then gives the equilibrium level of quality in the three-stage game,
v ¼ v S ðN Þ. Note that N is independent of S, c, and e. This is due to
the simplifying assumption f ¼ e. With a more general function FðuÞ,
N could be positively or negatively related to S, c, and e. This as-
sumption allows us, however, to focus on the variables of interest,
that is, l, b, and s.
Comparative Statics
The key comparative statics results of this model, within the range
of parameter values for which a symmetric equilibrium with N b 2
and v > 1 exists, have been analyzed elsewhere (Symeonidis 2000c).
They can be briefly summarized as follows.
First, a fall in l (more intense short-run competition) unambig-
uously reduces quality, and hence advertising expenditure. This is
true whether N is kept fixed—as can be easily checked by differen-
tiating v S ðNÞ with respect to l—or is allowed to adjust to its equilib-
rium value according to equation (5.9). Thus more price competition
implies less nonprice competition. While this result may be spe-
cific to the linear demand model, it is precisely because of this result
that the possibility of an ambiguous effect of a change in l on N
arises.
Indeed, the second key result is that a fall in l may cause either an
increase or a decrease in N , depending on the values of b and s and
the initial value of l itself. The ambiguity of the competition effect on
concentration in this model stems from the fact that more intense
price competition reduces advertising expenditure: since a fall in l
causes both gross profit and sunk cost to decrease, for given N, it is
Price Competition, Advertising, and Market Structure 169
" #
q dA
qb dl l¼0
where
E5 ¼ Sð1 cÞ 2 =e;
E6 ¼ 2s þ bð4 sÞ,
E7 ¼ ðb 2Þ½ðb 2 þ 2sÞð2 sÞ þ 2ðb sÞ 2
are all positive expressions for b > 2, s A ð0; 2Þ. Moreover, it can be
checked from equations (5.8) and (5.9) that ðv Þ b2 jl¼0 ¼ E5 =E62 . Since
we require v > 1, we also have lnðE5 =E62 Þ > 0. Hence ðdA =dlÞjl¼0 is
decreasing in b. By continuity, dA =dl is decreasing in b in the
neighborhood of the Cournot-Nash equilibrium. r
I use sales revenue by UK firms, while in the latter case I use sales
revenue in the UK market.
The new variables in this chapter are advertising intensity and a
set of variables that capture differences across industries and over
time in the fraction of TV advertising in total advertising. To con-
struct these variables, systematic data on advertising expenditure
across a range of advertising-intensive industries were required.
These data are described in some detail below.
Advertising-sales ratios were constructed for 1954, 1958, 1963,
1968, and 1973. The first four of these years are the only ones during
the 1950s and the 1960s for which comparable sales revenue data
across industries are available at a fairly disaggregated level. On the
other hand, 1973 is the last year before the recession of the mid-
1970s, which seems to have caused a temporary but significant fall in
advertising expenditure in many industries. The data used in the
concentration regressions are, as in chapter 4, for 1958, 1963, 1968,
and 1975. Recall that 1954 and 1958 are ‘‘before’’ dates in the natural
experiment, 1963 is a year when we expect to see only the short-run
effects of competition, 1968 is a year when much of the long-run ef-
fect should have been realized, and by 1973 or 1975 the full long-run
effects of competition had certainly been realized.
13. Other selling expenses include ‘‘direct selling activities’’ such as visits by sales-
persons, which are common in producer-good industries, and ‘‘promotional activities’’
such as coupons to consumers and special terms to retailers, which are common in
consumer-good industries.
Price Competition, Advertising, and Market Structure 175
14. A minor problem was that while the Census sales revenue data are for plants (for
1968 and 1973) or firms (for 1954, 1958, and 1963) employing at least twenty-five per-
sons, the import and export statistics cover all firms. Since very small firms often pro-
duce ancillary rather than core industry products and may spend little on advertising
or advertise only locally, the exclusion of very small firms (or plants) is a minor issue
for many industries, especially when imports and exports are small. In cases where
imports or exports were significant, or very small UK producers accounted for a con-
siderable fraction of total sales, an approximate adjustment of the sales figures was
made using published estimates of sales by very small UK firms, often available only
at a higher level of aggregation.
176 Chapter 5
15. The reason for applying a uniform adjustment of 10% to all press advertising fig-
ures is as follows. In the case of press advertising, a large part of the difference be-
tween Legion/MEAL and Advertising Association data relates to local advertising by
very small firms and to industrial or retail advertising or advertising of services, all of
which were not relevant for my present purposes. Also, much of the improvement in
Legion press coverage during the 1950s and the 1960s was primarily in these sectors.
In fact, I assumed that the improvement was wholly in these sectors, which in turn
implied that the extent of underreporting of media advertising by larger firms in con-
sumer-good industries was fairly uniform over the period. Hence I applied a uniform
upward adjustment to all Legion press figures. The choice of the 10% figure was based
on a comparison of Legion with Advertising Association aggregate figures for 1968:
the Legion press aggregate was 84% of the Advertising Association aggregate, and
much (but not all) of this must reflect incomplete coverage of press advertising by
larger firms in consumer-good industries.
178 Chapter 5
Table 5.1
Changes in C5 and ADS, 1958–1968
DC5 > 0 DC5 > 0 DC5 < 0 DC5 < 0
D ADS > 0 D ADS < 0 D ADS > 0 D ADS < 0 Total
No. of industries with 1 4 0 1 6
average ADS > 1%
and CHANGE ¼ 1
No. of industries with 7 8 2 6 23
average ADS > 1%
and CHANGE ¼ 0
No. of industries with 1 4 0 0 5
average ADS > 2%
and CHANGE ¼ 1
No. of industries with 4 4 2 6 16
average ADS > 2%
and CHANGE ¼ 0
Notes: The figures in the first two rows are based on twenty-nine advertising-intensive
industries with available comparable data for C5 and ADS for both 1958 and 1968. The
figures in the last two rows are based on twenty-one high-advertising industries with
available comparable data for C5 and ADS for both 1958 and 1968.
Econometric Methodology
16. Even in those cases, the comparability of the figures would be affected only if ad-
vertising intensity differed considerably between UK products and imported goods or
export intensity varied considerably across leading UK firms in each industry.
180 Chapter 5
17. In a few cases where there was evidence of a substantial change of this fraction
after 1968 or of an unstable allocation between press and TV advertising in general, the
average or typical value for the late 1960s and early 1970s was used instead of the 1968
figure.
Price Competition, Advertising, and Market Structure 183
concern in this case. On the other hand, TVCLASS is less likely than
TVINT to be subject to measurement error with respect to the theo-
retical variable e, and this is particularly important in the concentra-
tion regressions because the variation in e in the sample used in these
regressions is relatively low.18
18. The reason for this is twofold. First, many industries in the samples used in the
concentration regressions, especially in the sample of high-advertising industries, have
a high fraction of TV in total advertising. Second, the first year in these samples is
1958, when TV advertising was already quite established: about 25% of total adver-
tising in 1958 was TV advertising, as compared to zero in 1954 and about 36% in the
late 1960s and early 1970s.
19. Unlike media advertising, promotional activities such as gifts and coupons to
consumers, and financial inducements to distributors affect price competition rather
than increasing perceived quality or providing information about a product.
184 Chapter 5
crisis. Note that time dummies can control for these effects only im-
perfectly, since the impact of changes in distribution patterns on ad-
vertising intensity must have varied across industries. This problem
is relevant mostly for the last two years of the panel used for the
advertising regressions, 1968 and 1973. Other factors that may be
associated with the fall in aggregate advertising intensity in the late
1960s and early 1970s include the operation of prices and incomes
policy, the acceleration of the inflation rate, and the increase in import
penetration.
An additional source of noise in the data, this time during 1954–
1958, is the fact that in many respects British industry did not fully
emerge from the implications of wartime controls until the second
half of the 1950s. Advertising was still low in 1954 compared with
the late 1930s, and it was growing rapidly during the mid-1950s even
before the advent of commercial TV. This evolution can be captured
only partly by a year dummy for 1954, since it was not uniform
across industries.
Table 5.2
The samples used in the concentration regressions: advertising-intensive industries
and high-advertising industries
No. of No. of No. of No. of
indus- indus- indus- indus-
tries with tries with tries with tries with
CHANGE ¼ 1 CHANGE ¼ 0 CHANGE ¼ 1 CHANGE ¼ 0
and average and average and average and average
Data available for ADS > 1% ADS > 1% ADS > 2% ADS > 2%
All four years 10 30 5 16
1958, 1963, 1968 0 1 0 1
1963, 1968, 1975 0 11 0 5
1958, 1963 0 3 0 0
1963, 1968 0 1 0 1
Total 10 46 5 23
20. I have also run regressions using more balanced samples, including only indus-
tries with a full set of observations for at least the core period 1958–1968. The results
were very similar to those reported below. Recall that the same was true for exogenous
sunk cost industries. See Symeonidis (2000a) for a discussion of these results (with the
minor qualification that data for 1977, instead of 1975, are used in that paper).
186 Chapter 5
Table 5.3
Concentration in 1958 and competition regime in advertising-intensive and high-
advertising industries
Mean C5 in 1958
(st. deviation of C5)
Advertising-intensive industries
Industries with change of competition regime after 0.737 (0.131)
1958 (n ¼ 10)
Industries without change of competition regime after 0.609 (0.270)
1958 (n ¼ 34)
High-advertising industries
Industries with change of competition regime after 0.707 (0.158)
1958 (n ¼ 5)
Industries without change of competition regime after 0.686 (0.231)
1958 (n ¼ 17)
Notes: The figures are based on industries with available data for 1958; n indicates the
number of industries.
Table 5.4
Average change in C5, 1958–1968 and 1958–1975, in advertising-intensive and high-
advertising industries
DC5 1958–1968 DC5 1958–1975
Advertising-intensive industries
Industries with CHANGE ¼ 1 0.100 0.130
(0.075) (0.101)
n ¼ 10 n ¼ 10
Industries with CHANGE ¼ 0 0.049 0.038
(0.122) (0.139)
n ¼ 31 n ¼ 30
High-advertising industries
Industries with CHANGE ¼ 1 0.136 0.150
(0.082) (0.123)
n¼5 n¼5
Industries with CHANGE ¼ 0 0.047 0.031
(0.098) (0.113)
n ¼ 17 n ¼ 16
Notes: The figures in the first column are based on industries with available observa-
tions for both 1958 and 1968. The figures in the second column are based on industries
with available observations for both 1958 and 1975. The figures in parentheses are
standard deviations. n indicates the number of industries.
Price Competition, Advertising, and Market Structure 187
CHANGE ¼ 0 and available data for 1958. As will be seen below, this
class is the crucial one for testing the theoretical predictions devel-
oped in this chapter.
On the other hand, table 5.4 shows that the average increase in C5
after 1958 was much larger for industries affected by the legislation
than for industries not affected, whatever the sample used. For in-
stance, for the ADS > 2% sample, the average change in C5 between
1958 and 1968 was 13.6 percentage points for industries with a
change in competition regime and only 4.7 percentage points for the
control group. For the ADS > 1% sample, the respective numbers for
the average change in C5 over 1958–1968 are 10.0 percentage points
and 4.9 percentage points. Similar differences are observed for the
1958–1975 period. These differences are so large that they can be
expected to persist even when controlling for other variables.
A comparison between industries that experienced a large increase
in C5 over the period 1958–1975 and industries that experienced a
large decrease in C5 over the same period is also very informative.
The comparison is limited to the forty advertising-intensive industries
with available data for both these years. Of the five industries with the
largest rise in C5, two (biscuits for human consumption; washing
machines, electrically operated) are industries with CHANGE ¼ 1,
and three (corsets and brassieres; television receiving sets; knitted,
netted, or crocheted goods: hosiery) have CHANGE ¼ 0. However,
all five industries with the largest fall in C5 (tufted carpets, carpeting,
and carpet floor rugs; vegetables, quick frozen; fish and fish prod-
ucts, quick frozen; cutlery; powered industrial trucks and industrial
tractors) have CHANGE ¼ 0.
21. The random-effects model gives similar results with respect to the market size
variable and the competition variables, but it is always rejected by the Hausman test
in favor of the fixed-effects model.
Price Competition, Advertising, and Market Structure 189
dence of any difference in the timing of the impact of the 1956 Act on
concentration across classes of industries. Second, a few agreements
in advertising-intensive industries continued until late into the 1960s
before being abandoned. Third, the effect of the Act after 1968 in
advertising-intensive industries appears to be more pronounced in
regressions using logit C5, and it is precisely in these regressions that
heteroskedasticity is also more pronounced, which suggests that
these particular results may be disproportionately influenced by a
few industries with very high concentration. Fourth, the rise in con-
centration in previously cartelized industries after 1968 does not
appear particularly pronounced when one looks at the descriptive
statistics of table 5.4. And, finally, the descriptive statistics in table
5.4, especially when read in conjunction with those in table 4.3, sug-
gest that an important factor in the estimated impact of the 1956 Act
over time was the evolution of industries not affected by the legisla-
tion. In particular, while the average C5 increased over 1968–1975 in
exogenous sunk cost industries without a change in competition re-
gime, it fell slightly in the equivalent group of advertising-intensive
industries.
All this is not to say that any estimated effect of the 1956 Act on
concentration after 1968 in advertising-intensive industries is neces-
sarily more apparent than real. One potentially relevant factor in this
respect is the widespread use of resale price maintenance in these
industries, in contrast with exogenous sunk cost industries. It may
be the case that resale price maintenance delayed or hindered the
emergence of fully effective price competition among manufacturers
in some previously cartelized industries. Its abolition in the second
half of the 1960s may therefore be one factor behind the rise in con-
centration after 1968 in these industries.
Another potential factor is the evolution of advertising. A fall in
advertising expenditure in previously cartelized industries may have
occurred and helped absorb part of the pressure on net profits for
several years after 1958. Once advertising had fallen to very low
levels, however, it could not fall any farther, which may be why we
see an effect of the Act on concentration continuing after 1968 in
some advertising-intensive industries (assuming that competition
further intensified after that date).
A final interpretation is that the changes in the pattern of distri-
bution and the consequent fall in advertising across manufacturing
industries after the mid-1960s created favorable circumstances for a
Table 5.5
190
Regression results for concentration for industries with average ADS > 1% (fixed-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.102 — 0.102 — 0.71 — 0.73 —
(0.018) (0.018) (0.12) (0.13)
ln DS — 0.085 — 0.085 — 0.55 — 0.57
(0.022) (0.022) (0.14) (0.15)
ln K=N 0.001 0.004 — — 0.28 0.31 — —
(0.029) (0.031) (0.22) (0.24)
ln K=L — — 0.012 0.011 — — 0.34 0.33
(0.032) (0.035) (0.27) (0.28)
TVCLASS 0.023 0.013 0.024 0.013 0.54 0.44 0.52 0.42
(0.036) (0.039) (0.036) (0.039) (0.26) (0.28) (0.26) (0.28)
Y63 0.050 0.049 0.053 0.050 0.43 0.41 0.42 0.39
(0.020) (0.021) (0.018) (0.019) (0.13) (0.14) (0.12) (0.13)
Y68 0.095 0.094 0.102 0.097 0.81 0.79 0.81 0.76
(0.029) (0.031) (0.027) (0.029) (0.22) (0.23) (0.21) (0.22)
Y75 0.114 0.114 0.125 0.120 0.99 0.97 1.01 0.95
(0.040) (0.044) (0.037) (0.042) (0.28) (0.30) (0.29) (0.31)
CHANGE Y63 0.025 0.027 0.025 0.027 0.16 0.17 0.16 0.18
(0.031) (0.033) (0.031) (0.032) (0.19) (0.20) (0.19) (0.20)
Chapter 5
CHANGE Y68 0.039 0.046 0.038 0.046 0.29 0.34 0.29 0.34
(0.032) (0.032) (0.032) (0.032) (0.21) (0.21) (0.21) (0.21)
CHANGE Y75 0.064 0.065 0.063 0.064 0.67 0.69 0.67 0.69
191
192 Chapter 5
The evidence reported in table 5.5 suggests that the long-run effect of
competition and market size on concentration is not significantly
different in advertising-intensive industries than in exogenous sunk
cost industries. This may be because the 1% cutoff point is too low. In
particular, the results in table 5.5 may be driven by industries with
medium advertising intensity, and in these industries a positive
competition effect on concentration may be due to the fact that ad-
vertising cannot fall by much anyway. To obtain more decisive evi-
dence on the present theory, it is therefore necessary to examine the
sample of industries with typical ADS > 2%, that is, industries where
endogenous sunk costs represent a more significant fraction of total
sunk costs.
Results for the ADS > 2% sample using a fixed-effects specification
are presented in table 5.6, and results for the random-effects model,
using generalised least squares estimation, are reported in table 5.7.
The Hausman test is inconclusive in this instance. In any case, the
two sets of results are similar, except for the coefficients on the setup
cost proxies, which are statistically significant in the random-effects
model only.
Price Competition, Advertising, and Market Structure 193
The first thing to note is that the market size coefficients in these
regressions are nowhere statistically significant, suggesting that the
concentration-market size relationship breaks down in industries
with high advertising intensity. This is consistent with Sutton’s (1991)
robust prediction that concentration may either rise or fall as market
size increases in industries with significant endogenous sunk costs. 22
On the other hand, the competition effect is statistically significant
and large, although its magnitude should be interpreted with caution
since the number of industries with a change of competition regime
in the ADS > 2% sample is only five (and none of these has a very
high typical ADS).23 In any case, the 95% confidence intervals for
the coefficients on CHANGE Y68 or on CHANGE Y75 in these and
the previous regression results largely overlap, so there is no strong
evidence of any difference in the magnitude of the long-run competi-
tion effect on concentration across classes of industries.
To check whether any of the five industries with a change of re-
gime has a disproportionate influence on the results, I reestimated
the model, excluding each of the five industries in turn. The coeffi-
cient on CHANGE Y75 remained significant at the 5% or the 10%
level in all cases.
Since TVCLASS is a somewhat crude proxy, it may also be inter-
esting to see what happens if it is dropped altogether. In fact, the
results are very similar to those in tables 5.6 and 5.7. Finally, using
TVINT instead of TVCLASS again produces results similar to those
reported here. The absence of any significant effect of TVCLASS or
the market size variables on concentration is not inconsistent with
the theory of section 5.2: a rise in advertising effectiveness e or market
size S causes both advertising expenditure and gross profit to in-
crease, given the initial number of firms, so the effect on concentra-
tion is ambiguous. Only the joint outcome DA a 0 and DC b 0 is
ruled out. In fact, the lack of any systematic market size effect on
concentration is one of the predictions of the specific example pre-
sented in section 5.3.
A potential concern with the above results is that none of the
industries with a change in competition regime in the samples used
had an advertising-sales ratio higher than 5% in 1958, and many had
22. Other econometric studies that have confirmed this prediction include Sutton (1991,
chapter 5), Matraves (1992), Lyons and Matraves (1996), and Robinson and Chiang
(1996).
23. The five industries are biscuits, cocoa products, condensed milk, electric cookers
(stoves), and washing machines.
Table 5.6
194
Regression results for concentration for industries with average ADS > 2% (fixed-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.013 — 0.013 — 0.21 — 0.20 —
(0.028) (0.029) (0.24) (0.25)
ln DS — 0.037 — 0.037 — 0.09 — 0.08
(0.026) (0.026) (0.21) (0.20)
ln K=N 0.002 0.003 — — 0.05 0.01 — —
(0.059) (0.059) (0.32) (0.33)
ln K=L — — 0.013 0.018 — — 0.07 0.07
(0.069) (0.069) (0.43) (0.44)
TVCLASS 0.047 0.057 0.045 0.056 0.32 0.15 0.32 0.15
(0.057) (0.058) (0.057) (0.058) (0.49) (0.48) (0.49) (0.48)
Y63 0.022 0.018 0.026 0.021 0.31 0.26 0.31 0.24
(0.035) (0.035) (0.032) (0.032) (0.22) (0.21) (0.21) (0.21)
Y68 0.034 0.026 0.043 0.034 0.29 0.21 0.29 0.16
(0.056) (0.055) (0.052) (0.051) (0.32) (0.31) (0.33) (0.32)
Y75 0.007 0.007 0.023 0.006 0.20 0.05 0.19 0.02
(0.083) (0.081) (0.082) (0.079) (0.47) (0.45) (0.52) (0.50)
CHANGE Y63 0.036 0.035 0.034 0.033 0.14 0.13 0.14 0.14
(0.051) (0.051) (0.051) (0.051) (0.28) (0.28) (0.28) (0.28)
Chapter 5
CHANGE Y68 0.086 0.083 0.083 0.080 0.61 0.61 0.61 0.63
(0.052) (0.052) (0.051) (0.051) (0.32) (0.30) (0.32) (0.31)
CHANGE Y75 0.124 0.127 0.120 0.124 0.96 1.01 0.96 1.03
195
Table 5.7
196
Regression results for concentration for industries with average ADS > 2% (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.019 — 0.002 — 0.03 — 0.10 —
(0.023) (0.023) (0.23) (0.24)
ln DS — 0.001 — 0.010 — 0.14 — 0.23
(0.021) (0.020) (0.19) (0.20)
ln K=N 0.067 0.057 — — 0.57 0.49 — —
(0.033) (0.034) (0.26) (0.26)
ln K=L — — 0.083 0.077 — — 0.70 0.66
(0.048) (0.048) (0.34) (0.34)
TVCLASS 0.064 0.072 0.060 0.067 0.04 0.04 0.07 0.001
(0.057) (0.055) (0.059) (0.057) (0.55) (0.52) (0.56) (0.53)
Y63 0.007 0.007 0.006 0.005 0.14 0.12 0.13 0.10
(0.029) (0.029) (0.029) (0.029) (0.20) (0.21) (0.20) (0.21)
Y68 0.005 0.005 0.008 0.010 0.15 0.17 0.17 0.21
(0.035) (0.034) (0.037) (0.037) (0.24) (0.23) (0.25) (0.25)
Y75 0.052 0.053 0.063 0.068 0.53 0.56 0.62 0.69
(0.048) (0.048) (0.055) (0.055) (0.33) (0.32) (0.38) (0.37)
CHANGE Y63 0.035 0.034 0.041 0.040 0.04 0.03 0.09 0.08
(0.051) (0.051) (0.048) (0.048) (0.32) (0.32) (0.29) (0.28)
Chapter 5
CHANGE Y68 0.088 0.087 0.094 0.093 0.59 0.56 0.64 0.62
(0.041) (0.042) (0.041) (0.042) (0.29) (0.29) (0.27) (0.28)
CHANGE Y75 0.125 0.127 0.136 0.137 1.00 1.01 1.09 1.10
197
198 Chapter 5
24. One reason why low-advertising industries were excluded from the present anal-
ysis is that any systematic effect of price competition on advertising will be more dif-
ficult to identify and also less relevant in industries where advertising is not a key
strategic variable. Moreover, large proportional changes in advertising intensity in
low-advertising industries may have an influence on the results which is dispropor-
tionate to their economic significance. Finally, note that all the theoretical predictions
in this chapter are for advertising-intensive industries.
Price Competition, Advertising, and Market Structure 199
Table 5.8
The samples used in the advertising regressions: advertising-intensive industries and
high-advertising industries
No. of No. of No. of No. of
indus- indus- indus- indus-
tries with tries with tries with tries with
CHANGE ¼ 1 CHANGE ¼ 0 CHANGE ¼ 1 CHANGE ¼ 0
and average and average and average and average
Data available for ADS > 1% ADS > 1% ADS > 2% ADS > 2%
All five years 12 31 8 23
1954, 1958, 1963, 1968 2 9 1 5
1958, 1963, 1968, 1973 1 0 1 0
1954, 1958, 1963 0 2 0 2
1958, 1963, 1968 1 2 1 2
1963, 1968, 1973 0 6 0 5
1958, 1963 0 2 0 2
1963, 1968 0 3 0 1
Total 16 55 11 40
2% over the three core years 1958, 1963, and 1968, was also con-
structed. Note that sample selection is essentially done here on the
basis of the average ADS—which is likely to be largely driven by ad-
vertising effectiveness, an exogenous variable—so there is no selec-
tion bias. The samples do not include industries with ambiguous
state of competition in 1958 (or in the 1960s and early 1970s). Nor do
they include industries with available observations only for 1954–
1958 or 1968–1973. However, they do include industries with avail-
able data for a subset of the period 1958–1968.25 In most cases
observations are available for all five years in the sample (1954, 1958,
1963, 1968 and 1973). The two samples include 306 and 219 obser-
vations, respectively. Table 5.8 gives details on their structure.26 The
full set of industries and information on key variables are given in
table B3 of appendix B.
Tables 5.9 and 5.10 present descriptive statistics on initial levels
and changes in the advertising-sales ratio, ADS, separately for in-
25. Industries with a switch of regime and with advertising or sales data available for
only a subset of the period 1958–1968 would have normally been excluded, but there
were no such cases in these samples.
26. I have also run regressions using more balanced samples, including only indus-
tries with a full set of observations for at least the core period 1958–1968. The results
were very similar to those reported below.
200 Chapter 5
Table 5.9
ADS, 1954 and 1958, and competition regime in advertising-intensive and high-
advertising industries
Mean ADS in 1954 Mean ADS in 1958
(st. deviation of ADS) (st. deviation of ADS)
Advertising-intensive industries
Industries with change of 0.025 (0.023) 0.035 (0.028)
competition regime after
1958 (n ¼ 14)
Industries without change 0.057 (0.067) 0.071 (0.076)
of competition regime
after 1958 (n ¼ 42)
High-advertising industries
Industries with change of 0.032 (0.027) 0.045 (0.032)
competition regime after
1958 (n ¼ 9)
Industries without change 0.076 (0.071) 0.093 (0.080)
of competition regime
after 1958 (n ¼ 30)
Notes: The figures are based on industries with available data for 1954 and 1958;
n indicates the number of industries.
201
202 Chapter 5
The econometric model that I will use is a panel data model with
industry-specific effects. The basic specification is
ln ADSit ¼ a i þ g1 ln SSDOMit þ g2 lnðK=LÞit þ g3 TVINTit
27. Moreover, of the five industries with the largest percentage rise in ADS between
1958 and 1968, only one (chocolate confectionery) has CHANGE ¼ 1, while four (razors
and blades; sanitary towels; ice cream and ice lollies (ice pops); margarine) have
CHANGE ¼ 0. On the other hand, three out of five industries with the largest percent-
age fall in ADS between 1958 and 1968 have CHANGE ¼ 1 (typewriters; domestic
refrigerators and deep freezers; lead pencils), and only two (baby carriages; toilet
brushes) have CHANGE ¼ 0.
Price Competition, Advertising, and Market Structure 203
28. The reasons for looking here at proportional rather than absolute changes in ADS
are the same as the reasons for using ln ADS rather than the untransformed ADS as
dependent variable in the regressions.
Price Competition, Advertising, and Market Structure 205
Tables 5.11 and 5.12 contain regression results for the two samples.
Since industries with a change in competition regime had, on aver-
age, a much lower advertising-sales ratio throughout the period than
industries in the control group, the competition variable CHANGE
is probably correlated with the industry-specific effects in the panel
data model described above. To obtain consistent estimates, fixed-
effects estimation was used throughout.
Preliminary tests suggested the presence of a clear outlier, namely, the
1963 observation for the washing machine industry. The advertising-
sales ratio reached a record 12.6% in 1963 in this industry, compared
with 4.4% in 1958 and 2.9% in 1968. As described in detail in the
following section, this was due to very special circumstances in that
industry, namely, the temporary entry of firms that used press ad-
vertising as part of their process of distribution. These special cir-
cumstances and the fact that this type of advertising was definitely
nonstandard justify dropping this observation from the sample. In
any case, this affects only the coefficient on CHANGE Y63, which is
smaller in absolute value and statistical significance when the obser-
vation is included; the other results are not significantly affected.
The two tables report results for the basic model as well as for
several alternative specifications. Thus, I report regression results
using TVINT, and also using TVTOT. Recall that the main concern
with TVTOT is its potential endogeneity. However, this may not be a
serious problem after all: in preliminary regressions using TVINT as
an instrument for TVTOT, the Wu-Hausman test always rejected the
null hypothesis of a significant effect on the estimated coefficients of
the potential endogeneity of TVTOT even at the 20% significance
level. Moreover, since both TVINT and TVTOT are at best imperfect
controls for the effect of TV advertising, I also report results for a
specification that does not include either of these variables. Finally,
I have also experimented with a specification including the interac-
tion variables ln SSDOM Y54, ln SSDOM Y63, ln SSDOM Y68,
and ln SSDOM Y73 in order to check whether the negative link
Table 5.11
Regression results for advertising intensity for industries with average ADS > 1%
(fixed-effects estimation)
Dependent variable: ln ADS
ln SSDOM 0.21 0.33 0.22 0.34 0.21 0.34
(0.10) (0.10) (0.10) (0.10) (0.11) (0.11)
ln SSDOM Y54 — 0.05 — 0.06 — 0.06
(0.06) (0.06) (0.07)
ln SSDOM Y63 — 0.10 — 0.09 — 0.11
(0.05) (0.05) (0.05)
ln SSDOM Y68 — 0.17 — 0.17 — 0.19
(0.05) (0.05) (0.05)
ln SSDOM Y73 — 0.21 — 0.21 — 0.22
(0.06) (0.06) (0.06)
ln K=L 0.11 0.13 0.04 0.19 0.01 0.25
(0.18) (0.18) (0.17) (0.17) (0.17) (0.17)
TVINT 2.02 1.73 — — — —
(0.69) (0.64)
TVTOT — — 0.62 0.59 — —
(0.16) (0.15)
Y54 0.09 0.34 0.13 0.43 0.35 0.20
(0.12) (0.63) (0.10) (0.62) (0.10) (0.66)
Y63 0.04 0.99 0.09 0.90 0.09 1.07
(0.08) (0.50) (0.08) (0.49) (0.08) (0.49)
Y68 0.16 1.94 0.17 1.97 0.11 2.09
(0.13) (0.53) (0.12) (0.51) (0.12) (0.54)
Y73 0.50 2.81 0.55 2.89 0.48 2.87
(0.19) (0.70) (0.18) (0.69) (0.18) (0.73)
CHANGE Y54 0.19 0.14 0.21 0.16 0.14 0.09
(0.18) (0.16) (0.18) (0.15) (0.18) (0.15)
CHANGE Y63 0.14 0.13 0.17 0.16 0.16 0.14
(0.12) (0.11) (0.12) (0.11) (0.12) (0.11)
CHANGE Y68 0.30 0.27 0.33 0.30 0.32 0.28
(0.13) (0.12) (0.13) (0.11) (0.14) (0.12)
CHANGE Y73 0.22 0.24 0.26 0.28 0.21 0.23
(0.20) (0.18) (0.20) (0.19) (0.21) (0.19)
R2 0.37 0.43 0.38 0.45 0.33 0.40
R 2LSDV 0.89 0.90 0.90 0.91 0.89 0.90
No. of industries 71 71 71 71 71 71
No. of industries 16 16 16 16 16 16
with CHANGE ¼ 1
No. of observations 305 305 305 305 305 305
Note: Heteroskedasticity-consistent standard errors in parentheses.
Table 5.12
Regression results for advertising intensity for industries with average ADS > 2%
(fixed-effects estimation)
Dependent variable: ln ADS
ln SSDOM 0.24 0.38 0.24 0.37 0.26 0.39
(0.13) (0.13) (0.13) (0.13) (0.13) (0.13)
ln SSDOM Y54 — 0.12 — 0.11 — 0.13
(0.07) (0.07) (0.07)
ln SSDOM Y63 — 0.15 — 0.14 — 0.16
(0.05) (0.06) (0.05)
ln SSDOM Y68 — 0.19 — 0.19 — 0.20
(0.06) (0.06) (0.06)
ln SSDOM Y73 — 0.25 — 0.25 — 0.25
(0.07) (0.07) (0.07)
ln K=L 0.07 0.31 0.12 0.32 0.12 0.34
(0.22) (0.21) (0.22) (0.21) (0.21) (0.20)
TVINT 1.34 0.50 — — — —
(0.80) (0.73)
TVTOT — — 0.42 0.31 — —
(0.18) (0.16)
Y54 0.11 0.83 0.13 0.84 0.31 0.83
(0.16) (0.68) (0.13) (0.68) (0.10) (0.67)
Y63 0.03 1.43 0.06 1.34 0.08 1.46
(0.11) (0.56) (0.10) (0.57) (0.10) (0.56)
Y68 0.19 2.09 0.20 2.15 0.12 2.14
(0.18) (0.65) (0.17) (0.65) (0.16) (0.65)
Y73 0.54 3.11 0.59 3.19 0.48 3.14
(0.25) (0.71) (0.25) (0.72) (0.23) (0.71)
CHANGE Y54 0.25 0.09 0.28 0.13 0.24 0.08
(0.23) (0.15) (0.23) (0.15) (0.23) (0.14)
CHANGE Y63 0.12 0.16 0.14 0.17 0.12 0.16
(0.15) (0.12) (0.15) (0.13) (0.15) (0.12)
CHANGE Y68 0.32 0.36 0.35 0.38 0.33 0.36
(0.17) (0.14) (0.17) (0.13) (0.17) (0.14)
CHANGE Y73 0.20 0.28 0.24 0.31 0.20 0.28
(0.24) (0.20) (0.25) (0.20) (0.25) (0.20)
R2 0.36 0.46 0.37 0.47 0.34 0.45
R 2LSDV 0.86 0.88 0.87 0.88 0.86 0.88
No. of industries 51 51 51 51 51 51
No. of industries 11 11 11 11 11 11
with CHANGE ¼ 1
No. of observations 218 218 218 218 218 218
Note: Heteroskedasticity-consistent standard errors in parentheses.
208 Chapter 5
How should we interpret the results on the effect of the 1956 cartel
law on advertising intensity? As we have seen, the coefficient on
CHANGE Y68 is negative and statistically significant, while the
coefficient on CHANGE Y73 is negative but not statistically signi-
ficant. I would argue that the evidence is consistent with the view
that the intensification of price competition following the 1956 Act
probably had a negative effect on advertising intensity across a range
of UK industries during the 1960s. There are three different argu-
ments to support this view.
29. I also tried replacing ln K=L with ln K=N as my setup cost proxy. The results were
very similar to those presented here. They have been omitted, since this variable is not
directly relevant in the advertising regressions anyway.
Price Competition, Advertising, and Market Structure 209
30. The advertising data for the whole of this section were taken from the Statistical
Review of Press Advertising, the Statistical Review of Independent TV Advertising, the Sta-
tistical Review of Press and TV Advertising, and the MEAL Monthly Digest of Advertising
Expenditure. The figures were adjusted to take into account press undercoverage, dis-
counts for TV advertising, and production costs of advertisements. Advertising by
accredited dealers is included, together with advertising by manufacturers and
importers, but advertising by the Electricity Council and Electricity Boards (which
have been acting as retailers in the market) for own-brand appliances is not. Data on
sales revenue of UK producers, the five-firm concentration ratio, and firm numbers
were obtained from the Census of Production and Business Monitor, PQ series: Quar-
terly Statistics for various years, and relate to complete machines only. Some data on
market shares were also taken from market research sources. Import and export data
for various years are from the Annual Statement of the Trade of the United Kingdom.
214 Chapter 5
31. The rise of the five-firm sales concentration ratio from 76% in 1958 to 85% in 1963
was mainly due to the 15% market share of Rolls Razor (cf. Hart et al. 1973), and
should not conceal the fact that the structure of the industry remained too fragmented
for firms to be able to cover their sunk costs.
32. Other factors may also have played a role in the restructuring of the industry, such
as the fall in demand between 1963 and 1968, although it must be noted that adver-
tising expenditure fell by much more than sales revenue over that period.
216 Chapter 5
producer price index of the industry, which fell by more than 25%
between 1958 and 1963, while the producer price index for all man-
ufacturing increased by 8% over the same period.
Changes in distribution in the mid- and late 1960s, including the
growth of retailers’ own brands (Baden-Fuller 1980) may have put
further pressure on profit margins. An additional major influence on
competitive conditions in the industry since the mid-1960s was the
growth of imports. While in the 1950s these were not significant, in
1968 they accounted for more than a quarter of total manufacturers’
and importers’ sales revenue in the UK market. The producer price
index of the industry remained unchanged between 1963 and 1968,
at a time when the producer price index for all manufacturing
increased by about 11.5%.
Nonprice competition was an important element of manufacturers’
strategies in the 1950s, but its significance diminished over the fol-
lowing decade. Advertising expenditure in the press and on TV was
3.7% of total manufacturers’ and importers’ sales revenue in the do-
mestic market in 1958 and remained high during the early 1960s. The
most important advertisers throughout this period were Electrolux,
Frigidaire, and, until its closure in 1964, Pressed Steel. These three firms
were consistently among the four largest producers of refrigerators in
the 1950s. Note that part of the reason for the high advertising-sales
ratio of the industry in the early 1960s was a 1962 deal between Rolls
Razor and Pressed Steel, whereby Rolls Razor would market Pressed
Steel refrigerators. The venture accounted for more than a quarter of
the industry’s advertising expenditure in 1963, but for only about 7%
of sales, and Pressed Steel eventually ceased production. Thus it
seems reasonable to regard the high advertising of the early 1960s as
a short-run disequilibrium response to the new competitive regime
in the industry.
This interpretation is consistent with the fact that the industry ad-
vertising-sales ratio collapsed after the exit of Pressed Steel in 1964,
reaching 1.1% in 1968 and even lower levels in the 1970s.33 One rea-
son for this large fall in advertising intensity was the intensification
of price competition. Another reason must have been the negative
effect of the changes in distribution and the growth of retailers’ own
brands on the effectiveness of brand advertising by manufacturers.
34. The effect of market growth is to create a ‘‘bias’’ toward a rise in firm numbers and
a fall in concentration to the extent that the exogenous component of sunk costs is
significant. However, this may not be true in a high-advertising industry, since market
growth will then induce an escalation of advertising expenditure, all else being equal
(see Sutton, 1991). Moreover, other factors may influence concentration upward and
firm numbers downward, as is likely to have happened in the present case—see the
discussion in chapter 4 on the various factors that contributed to the overall rise in
concentration in the UK during the 1960s.
35. These mergers had a much smaller impact on concentration in the washing
machine industry than in the refrigerator industry because of the low market share of
G.E.C. in washing machines and a 1965 agreement whereby A.E.I.-Hotpoint took over
English Electric’s washing machine division and English Electric took over A.E.I.-
Hotpoint’s stove division (see Hart et al. 1973).
220 Chapter 5
It should be pointed out that the concentration data are not strictly
comparable with the data on firm numbers described above. The for-
mer refer to sales of all electrical refrigerators in the UK market; the
latter refer to sales of mechanically operated refrigerators by UK pro-
ducers employing at least twenty-five persons. Nevertheless, both
pieces of evidence point in the same direction. Moreover, the stabil-
ity, or even increase, of firm numbers and the apparent fall in con-
centration in the refrigerator industry between 1958 and 1968 should
be seen in the context of the overall substantial fall in firm numbers
and the increase in concentration in British manufacturing as a whole
during the 1960s. As pointed out in chapter 4, Hart and Clarke (1980)
report a rise of eight percentage points in the average five-firm sales
concentration ratio among UK producers at the four-digit industry
level between 1958 and 1968. My own calculations suggest that most
of this can be attributed to a fall in firm numbers, since the number of
UK firms employing at least twenty-five persons fell during the same
period by about 25–30% in the average industry (see chapter 7 of this
book). Compared to this benchmark, the refrigerator industry expe-
rienced a relative rise in firm numbers and a relative fall in concen-
tration during 1958–1968, even if foreign firms are not taken into
account.
Interestingly, this is not the end of the story for the refrigerator
industry. Competitive pressure continued to intensify after 1968, as
shown by the fact that the producer price index of the industry
increased by only 9.5% over 1968–1973, while the producer price in-
dex for all manufacturing increased by 37% over the same period.
Several factors may have caused price competition to intensify fur-
ther, and one of them was certainly import penetration: imports
increased from 25% of total sales in the UK in 1968 to almost 50% in
1973. Now advertising was already low in 1968, so it could not fall
much farther. Proposition 5.2 would then predict that during 1968–
1973 the number of firms should have fallen, everything else being
equal, since the pressure on margins could no longer be offset by any
significant reduction in sunk costs. This is indeed what happened:
according to Business Monitors, there were only six UK producers of
all types of domestic refrigerators still in operation in 1972–1973.
In summary, the experience of the refrigerator industry during
1958–1968 shows that a negative competition effect on concentration
is not just a theoretical oddity. This was a high-advertising industry
that experienced a significant increase in the intensity of price com-
Price Competition, Advertising, and Market Structure 221
falls little or not at all. Taken together, the theoretical results indicate
that a positive effect of price competition on concentration is the
most likely outcome in advertising-intensive industries, since the re-
verse can occur only in the event of a significant fall in advertising
expenditure.
In the second part of this chapter, the theory was tested against
econometric and case-study evidence on the evolution of advertising-
intensive manufacturing industries following the introduction of
cartel legislation in Britain in the late 1950s. The econometric results
indicate that the intensification of price competition following the
introduction of the legislation caused, on the whole, a rise in con-
centration of about six percentage points in advertising-intensive
industries, and probably also a fall in advertising intensity on the
order of 20–25%. The case-study evidence confirms that a negative
effect of price competition on concentration cannot be ruled out in
high-advertising industries. However, a necessary condition for this
to occur is that tougher price competition leads to a significant fall in
advertising intensity. On the other hand, if advertising falls little or
not at all when price competition intensifies, then concentration will
rise, which is consistent with the theoretical predictions as well as
with the econometric results.
6 Price Competition,
Innovation, and Market
Structure in R&D-
Intensive Industries
6.1 Introduction
A Long-Standing Debate
1. A common assumption in this context is that firms cannot collude in the long-run
variable, whatever their short-run conduct. This is often justified in theoretical studies
by a reference to the fact that long detection and retaliation lags hinder the stability of
collusion. Since long-term decisions take time to implement, the reaction to rivals’ be-
havior is relatively slow (i.e., there are relatively long retaliation lags); hence collusion
in long-run variables will be relatively difficult to achieve. In the case of R&D, an ad-
ditional argument is that deviations from agreed levels of R&D expenditure are diffi-
cult to observe, at least as long as R&D cooperation does not take the form of a joint
venture; in other words, detection is also problematic, and this further hinders collu-
sion on R&D.
Price Competition, Innovation, and Market Structure 225
(see Utterbach and Suarez 1993, Klepper 1996, Klepper and Simons
1997, Sutton 1996, 1998, among others).2
There have been numerous attempts to address one or the other
of these issues in econometric work. Thus some authors have ex-
amined the determinants of innovation using measures of compe-
titive pressure or market power other than market structure, such
as import penetration; some of this work will be reviewed below.
Others have tried to address the endogeneity issue by using instru-
mental variable techniques or by estimating simultaneous equation
systems, with rather mixed results (see, e.g., Farber 1981, Lunn 1986,
Levin and Reiss 1988).3 However, none of these studies has explicitly
modeled innovative output and concentration as jointly determined
by some exogenous measure of competition (as well as by other
variables). This is the approach that will be taken here.
Before describing my methodology in more detail, I will briefly
discuss a number of previous empirical studies of the determinants
of R&D or innovation that have used measures of market power
other than market structure. Taken as a whole, these studies provide
little evidence of a negative effect of competition on innovation, and
only weak evidence of a positive effect.
This is certainly the conclusion from a comparison of two studies
of the effects of import competition on R&D or innovation. The first
is a detailed study of R&D-spending reactions of US firms to in-
creased high-technology import competition over the period 1971–
1987, carried out by Scherer and associates (see Scherer 1992, Scherer
and Huh 1992). The authors found that the response varied consid-
erably across firms, and that most of this variation was unsystematic.
2. This recent literature consists of two main strands. One line of research has focused
on developing stylized theories of the life cycle of R&D-intensive industries and test-
ing them against evidence from industry case studies. See Klepper and Simons (1997)
for a comparison of different theories. A second approach has questioned the general-
ity of any particular evolutionary path, and has focused instead on a small number of
mechanisms that are likely to work across a broad range of industries. This is the
approach of Sutton (1996, 1998).
3. Innovation may affect market structure directly or indirectly, and in various and
contradictory ways. For instance, it may increase or decrease the minimum efficient
scale of production, and it may facilitate or hinder the entry of new firms. So it is per-
haps not surprising that the econometric literature on this issue has been inconclusive.
For instance, while Farber (1981) and Levin and Reiss (1988) found a positive effect of
R&D on concentration, Mukhopadhaya (1985) found a negative effect of R&D on con-
centration change, and Geroski and Pomroy (1990) found a negative effect of innova-
tion counts on concentration change.
Price Competition, Innovation, and Market Structure 227
4. The structure of the model analyzed in this section is similar to the structure of the
linear demand model of section 5.3. The competition measure t is the equivalent of
the inverse of the parameter l of the linear demand model. The innovation index I is
the equivalent of the quality index u, and the basic technological relationship between
Ii and Ri is similar to that between ui and Fi in the linear demand example. Finally,
assumptions R1–R3, made below, are all satisfied in the linear demand example. Of
course, the present model is much more general.
230 Chapter 6
The structure of this game reflects the fact that (1) decisions on
R&D expenditure involve significant sunk costs and (2) product or
process characteristics cannot be changed as easily or as quickly as
price choices, although they are still more flexible than entry deci-
sions. However, I abstract from issues of uncertainty regarding the
outcome of R&D projects, as do, in fact, many models within the
general class of ‘‘nontournament’’ models of R&D rivalry, of which
the present model is an example.5 The key underlying assumption of
these models, as opposed to the literature on patent races, is that
each firm can achieve product or process improvements through its
R&D expenditure without preventing other firms from obtaining
equivalent benefits from their own R&D.
7. To see this, note that the R&D of firm j has two opposing effects on the gross profit
of firm i. On the one hand, an increase in the R&D of firm j raises its innovation index,
and thus increases its market share and gross profit at the expense of firm i’s market
share and gross profit. This negative R&D externality is always present in the model.
On the other hand, when there are positive R&D externalities (spillovers), there is also
another effect: an increase in the R&D of firm j raises the innovation index of firm i, and
thus causes the gross profit of that firm to increase at the expense of firm j’s own gross
profit. The results of the model are not altered as long as this second effect does not
dominate the first (i.e., as long as spillovers are not too large). It is interesting in this
respect that Geroski (1994) found no evidence of significant R&D spillovers in British
industry during the period that I analyze in this book.
232 Chapter 6
8. The requirement that dRi =dIi < dpi =dIi at Ri ¼ 0 ensures that firms will choose to
carry out some R&D at equilibrium. The assumption of a continuous innovation pro-
duction function is, of course, a simplification. A more general formulation would
allow for the presence of indivisibilities in R&D projects, especially since in some
industries the absolute size of the typical R&D project may be large. However, the
existence of indivisibilities does not significantly alter the basic analysis of the incen-
tives to conduct R&D.
Price Competition, Innovation, and Market Structure 233
firm spends on R&D up to the point where the cost of an extra unit of
R&D is equal to the profit this creates in the third-stage subgame
through the corresponding rise in innovative output. At any sym-
metric equilibrium, we have Ii ¼ I (and hence Ri ¼ R), Ei, so the first-
order condition for firm i can be written as
d d
pi ðS; N; h; t; Ii ; ðIi ÞÞ ¼ Ri ðr; Ii Þ at Ii ¼ Ii ¼ I: ð6:1Þ
dIi dIi
9. It is, of course, true that firms may find it easier to collude on R&D when R&D co-
operation takes the form of a joint venture. However, as pointed out in chapter 2, there
was hardly any R&D cooperation of that form in British industries that practiced price-
fixing in the 1950s. More generally, allowing for collusion on R&D in the present
model would probably increase the likelihood that a breakdown of collusion (affecting
all choice variables) leads to a rise in R&D and innovations. A rise in R&D in this case
might also imply a rise in the net profit of each firm, for a given number of firms.
Proposition 6.1 below, which is a prediction on the joint effect of price competition on
innovations and market structure, would have to be modified to take that possibility
into account. If, however, a breakdown of collusion had no effect on R&D in the more
general model, proposition 6.1 should still hold unambiguously.
234 Chapter 6
net profit below zero, for given C and Ri ’s. For the zero-profit condi-
tion to be restored at a new symmetric equilibrium, net profit must
rise. Now suppose that the increase in t has also caused all firms to
spend more or exactly the same on R&D. This cannot increase net
profit because (1) an increase in own R&D has a stronger effect on
sunk cost than on gross profit when starting from a long-run equi-
librium with dpi =dRi ¼ 1, Ei (owing to the concavity of the gross
profit function with respect to Ri ), so it reduces gross profit; (2) if
own R&D does not change, it has no effect on profit; and (3) an in-
crease or no change in rival R&D cannot increase own profit. In these
circumstances, then, there is only one way net profit can increase,
through a rise in concentration.
Note that proposition 6.1 is a sufficient, but not necessary, condition
for a positive competition effect on concentration in R&D-intensive
industries. A positive competition effect on concentration can occur
even if the R&D expenditure and the innovative output of each firm
fall, provided that this fall is not too large and therefore does not
offset the direct negative effect of the rise in t on gross profit. Formal
theoretical results capturing this insight are easy to obtain, and they
are comparable to results obtained in section 5.2 of chapter 5 for
advertising-intensive industries. The key implication is that in R&D-
intensive industries, a nonpositive effect of competition on concen-
tration, although not impossible, may be unlikely.
Finally, note that proposition 6.1 is a prediction on the joint effect
of price competition on firm innovation and market structure. The
question arises as to whether anything can be said about the joint
effect of price competition on industry innovation and market struc-
ture. In particular, an interesting question is whether it is possible to
rule out the case DðN I Þ b 0 and DN b 0 ðDC a 0Þ, where DI ,
DN , and DC denote the change in I , N , and C , respectively. This
joint outcome can occur in one of two ways: either through DI b 0
and DN b 0; or through DI < 0 and DN b 0, with the rise in N
offsetting the fall in I , so that N I rises. The former case is ruled out
by proposition 6.1. The latter case cannot be ruled out on the basis of
the assumptions we have made, but it does seem very improbable.
To obtain DN b 0 following a rise in t, R needs to fall considerably;
if it falls by only a small amount, it will not offset the direct negative
effect of the rise in t on gross profit, so N will fall. But since a sig-
nificant fall in R is required for even a small rise (or no change) in
N , it is very difficult to obtain DðN I Þ b 0 together with DN b 0
ðDC a 0Þ following a rise in t.
Price Competition, Innovation, and Market Structure 237
Discussion
successful innovation.11 Now the larger this variance, the more likely
it is that an increase in market size will have a positive effect on
concentration, other things being equal. Of course, this is only one
potential mechanism through which market size may affect concen-
tration, and there are many others. The point is, however, that the
presence of potentially large asymmetries in the growth of firms in
R&D-intensive industries may cause a positive ‘‘bias’’ in the observed
market size effect on concentration relative to the benchmark predic-
tion for the symmetric case. Such a bias could be much more pro-
nounced in R&D-intensive industries than in other industries.
The theory should now be confronted with the empirical evidence.
The first question to be asked is what was the effect of the 1956 Act
on innovations in British industry. This is a question on which the
theory developed in this section is deliberately silent. Provided that
we observe a nonnegative effect, the next question will then be whether
concentration did indeed rise, as suggested by the present theory.
11. There is considerable evidence that the size distribution of the profit outcomes
from innovation is highly skewed. See, for instance, Scherer et al. (2000).
Price Competition, Innovation, and Market Structure 239
The innovations data used in this chapter come from a major survey
of significant innovations commercialized by UK firms during 1945–
1983. The survey was carried out by researchers in the Science Policy
Research Unit (SPRU) at the University of Sussex over a period of sev-
eral years and covers a total of 4,378 innovations. The data were col-
lected in three waves over a fifteen-year period: a first survey was
conducted in 1970 for the period 1945–1970 and identified 1,304 inno-
vations; a second, in 1980, covered the period 1970–1980 and identified
a further 848 innovations; and a third, carried out in 1983, covered
the whole period 1945–1983 and added another 2,226 innovations.
12. In particular, three groups of industries with typical RDS < 1% were initially
included in the enlarged sample: (1) industries in the nonferrous metal sector, whose
R&D-sales ratio was close to 1% for much of the period; (2) all industries with
RDS < 1% in the five core R&D-intensive sectors (chemicals, mechanical engineering,
instruments, electrical and electronic engineering, and vehicles); and (3) four more
industries with a large number of innovations in the database (iron and steel, glass
other than containers, paper, and plastic products). Some of these were later dropped
because of their ambiguous state of competition.
Price Competition, Innovation, and Market Structure 241
13. Each sector was covered by several experts who were drawn from research and
trade associations, academic institutions, government departments, trade and techni-
cal journals, individuals and consultants, as well as from firms.
242 Chapter 6
14. The low numbers of innovations during the early period may be partly due to the
difficulty of finding in 1970 experts who had been active during that period. And the
decline in innovation counts after 1980 may be due to the fact that this period was
covered in only one wave of data collection, while all the other years were covered in
two waves.
15. The share of innovations produced by larger firms is much lower than their share
of R&D expenditures, while the reverse is true for smaller firms (see Geroski 1994,
chapter 2). There may be various reasons for this, including underreporting of R&D by
small firms, overrepresentation of some low-concentration industries in the SPRU
database, failure to take into account links between small and large firms in the pro-
duction of innovations, and differences in the type of projects undertaken by larger
and smaller firms. Whatever the reasons, the SPRU data are certainly free from any
potential large-firm bias.
Price Competition, Innovation, and Market Structure 243
18. In cases where the innovating unit is a subsidiary of another firm, the data report
the principal industry classification both of the subsidiary and the parent firm.
Price Competition, Innovation, and Market Structure 245
19. The most common difficulty was the following. Suppose that the principal indus-
try classification of the innovating firm/unit is A, the industry classification of the first
user of the innovation is again A, and the industry classification of the innovation is B.
It is then not always easy to tell whether this is a process innovation in industry A or a
product innovation in industry B by a division of that particular innovating firm/unit.
Clearly, in such a case it is helpful to know the exact description of the innovation
and whether the firm in question has a plant in industry B. Information on the latter
question for the year 1968 is contained in a detailed Directory of Business published as
part of the Census of Production of that year. Although this directory is obviously less
useful for innovations introduced in the 1950s than for those introduced in the 1960s
and early 1970s, and its coverage is not complete, it often provided the key informa-
tion required for classifying particular innovations. In other cases, my classification
was based on more subjective criteria. For instance, cases in which industry A in the
above example was iron and steel and industry B was instruments were classified as
process innovations, since it seemed unlikely that steel firms would have divisions
producing instruments. On the other hand, cases where both A and B were electronic
engineering industries were usually classified as product innovations, since firms in
that sector tend to be diversified into related industries.
20. There is also a second problem with the way the innovations are assigned to the
various industries in the SPRU database, which applies especially to low-innovation
sectors, such as food and drink, textiles, and clothing. In these sectors, the innovations
are very often assigned to the sector rather than to any particular industry. In fact, they
are assigned to the first ‘‘minimum list heading’’ (MLH) industry in each of these sec-
tors, which is, of course, very misleading. In most of these cases, it was possible to
reclassify the innovations to the appropriate industries. In any case, since my empirical
analysis focuses on R&D-intensive or innovation-intensive industries, the sectors that
were subject to serious measurement errors of this kind are not included in the sam-
ples used in this chapter. Note that MLH 411, man-made fibers, which is the first MLH
industry in the textile sector, is R&D-intensive. This industry was also dropped from
my samples, however, because of ambiguous state of competition in the 1950s.
246 Chapter 6
Innovation Lags
Although the SPRU data record the number of innovations for every
year between 1945 and 1983, it is necessary to group the innovation
counts into somewhat longer periods, so that they can be matched
with Census-based figures on sales revenue and other variables that
are available only at roughly five-year intervals during the 1950s and
the 1960s. Using five-year periods to group the innovations is there-
fore an obvious choice, but this still leaves open the question of
exactly which periods to use.
Data on sales revenue, capital stock, employment, and so on are
available for 1954, 1958, 1963, 1968, and then annually from 1970 on-
ward. Data for 1953 are not available, but I constructed estimates for
1953 sales revenue on the assumption that the proportional change
of the deflated sales revenue in any given industry over 1953–1954
was equal to the average yearly rate of change of the deflated sales
revenue in that industry over 1954–1958. I used the same method to
construct estimates for the 1953 capital-labor ratio and the ratio of
capital stock to the number of plants in each industry. Now it is
reasonable to assume that the sales revenue in any given year t is a
measure of average market size between years t 2 and t þ 2. It is
Price Competition, Innovation, and Market Structure 247
The time periods chosen for grouping the innovations data in this
chapter are therefore 1952–1956, 1957–1961, 1962–1966, 1967–1971,
and 1972–1976. These were matched with data on sales revenue (and
capital intensity) for 1953 (my estimates), 1958, 1963, 1968, and 1973,
respectively. Note that these time periods are also very convenient in
the present case because focal-point years, such as 1955, 1960, 1965,
and so on are not borderline between any two consecutive periods.
This is important because the experts and the firms consulted for the
construction of the SPRU database may have used focal-point years
whenever they were uncertain about the exact date of the introduc-
tion of an innovation. The fact that focal-point years are close to the
midpoints of the time periods that I use here to group the innova-
tions should practically eliminate any errors resulting from imprecise
dating of innovations in the SPRU database.
These time periods are also convenient from the point of view of
the evolution of competition in British manufacturing. The number
of innovations commercialized in any given industry during 1952–
1956 was clearly determined by market conditions prior to the intro-
duction of the 1956 restrictive practices legislation. Innovative output
during the next period, 1957–1961, was driven by market conditions
during 1956–1960, which covers the last few years of cartelization as
well as the first two years after industries had started to formally
abandon price-fixing. But because of the fact that in many R&D-
intensive industries price-fixing agreements continued until the early
or even the mid-1960s, the period 1956–1960 can be seen, on the
whole, as a time when competition had not yet emerged to any sig-
nificant scale. The next period in the innovations data, 1962–1966, is
one in which we expect to see the short-run effect of competition on
innovation. In the final two periods, 1967–1971 and especially 1972–
1976, we almost certainly observe the long-run effect of competition.
While the assumption of a one-year innovation lag seems reason-
able and has been used in previous studies using the SPRU data (e.g.,
Blundell et al. 1995), I also experimented with an alternative set of
time periods for grouping innovation counts, chosen on the assump-
tion of an average two-year lag between the bulk of R&D spending
and the commercialization of innovations. This implied matching in-
novation periods 1953–1957, 1958–1962, 1963–1967, 1968–1972, and
1973–1977 with sales revenue data for 1953, 1958, 1963, 1968, and
1973, respectively. The choice of a two-year lag is consistent with the
evidence from a survey of R&D expenditure and innovations in a
Price Competition, Innovation, and Market Structure 249
sample of firms drawn among the 300 largest firms in British industry
in the early 1970s (Schott 1976). Note, however, that the study by
Schott focused on large or very large firms. On the other hand, a
large fraction of the innovations in the SPRU database are by small
firms, and the average R&D project of a small firm is likely to be less
expensive than that of a large firm. To the extent that this also implies
that smaller firms are generally involved in shorter-term projects than
larger firms, a one-year average innovation lag may be an appropriate
assumption for the SPRU data.
In any case, I chose to use the assumption of a one-year lag for
my basic specification of innovation counts (with dependent variable
INN1), but I also checked the robustness of the results under the
alternative assumption of a two-year average innovation lag (using
the alternative dependent variable INN2). Both sets of results will be
reported below, although they are, in fact, very similar.
22. Townsend et al. (1981) attribute this feature of the data to a variety of factors,
including a potential bias in the collected information. According to these authors, the
bias may have resulted from a mixture of technological nationalism, real difficulties in
distinguishing innovating firms and countries from imitating firms and countries, and
a high rate of import of foreign technology through mechanisms that are not easily
observable (such as imitation, reverse engineering, and inventing around patents), as
opposed to more measurable mechanisms (such as licensing, or transfers within mul-
tinational firms).
250 Chapter 6
23. It is also difficult to obtain any meaningful descriptive statistics on the joint evo-
lution of concentration and innovation counts in individual industries (such as those
reported for the joint evolution of advertising intensity and concentration in table 5.1).
There are two reasons for this. First, innovation counts tend to be volatile, so a rise or a
fall in innovation counts between the late 1950s and the late 1960s, say, in any given
industry may not in itself be very informative. Second, changes in innovative activity
are likely to be correlated with changes in sales, but it is not clear how to control for
this when focusing on the evolution of innovation counts in individual industries; for
instance, dividing the number of innovations by deflated sales revenue would give a
different picture than dividing it by the log of deflated sales revenue.
Price Competition, Innovation, and Market Structure 251
Table 6.1
Innovation counts in 1952–1956, 1957–1961, 1962–1966, 1967–1971 and 1972–1976
Mean (st. deviation) of INN1
1952–1956 1957–1961 1962–1966 1967–1971 1972–1976
industries with average RDS < 1% but with a relatively large num-
ber of innovations in the SPRU database. The full set of industries
and information on key variables are given in table B4 of appendix B.
Table 6.1 presents descriptive statistics for innovation counts sepa-
rately for industries with a change in competition regime and indus-
tries without such a change. The figures are based on the near-total
number of innovations introduced in British industry and recorded
in the SPRU database (i.e., excluding only a small number of inno-
vations produced by public research establishments or industrial
trade associations). Perhaps the most striking feature of the data is
the significant dissimilarity in initial conditions between the two
groups. In particular, the average number of innovations introduced
during 1952–1956 was 1.8 for R&D-intensive industries with a change
of competition regime in the 1960s and 5.4 for R&D-intensive indus-
tries without a subsequent change in regime (all of which were non-
collusive in the 1950s).
However, it would be very misleading to interpret this large dif-
ference in initial conditions as an indication of a negative effect of
collusion on innovative activity. An alternative interpretation, pro-
Price Competition, Innovation, and Market Structure 253
In other words, the time periods for grouping the innovations are
in this case 1953–1957, 1958–1962, 1963–1967, 1968–1972 and 1973–
1977. Finally, to construct INN3 all innovations classified in the SPRU
data as having originated outside the UK were excluded, and the
innovations were grouped into five-year periods according to a one-
year innovation lag.
The independent variables are defined as follows. ‘‘Sales’’ is either
total sales revenue by UK firms deflated by the general producer
price index ðSSÞ or total sales revenue deflated by an industry-specific
producer price index ðDSÞ. The level of aggregation for these data is
the same as that for the innovation measures and the competition
variable CHANGE—that is, sometimes it is the three-digit industry
level, sometimes it is the four-digit industry level, and sometimes it
is between the three-digit and the four-digit industry level. K=L is the
capital-labor ratio, defined at the three-digit industry level.24 Y58,
Y63, Y68, and Y73 are time dummies corresponding, respectively, to
time periods 1957–1961, 1962–1966, 1967–1971 and 1972–1976 when
INN1 or INN3 is used, and to 1958–1962, 1963–1967, 1968–1972, and
1973–1977 when INN2 is used.
Finally, the interaction terms should reveal any differences in the
evolution of innovation counts since the mid-1950s between indus-
tries with a change in competition regime ðCHANGE ¼ 1Þ and indus-
tries without such a change ðCHANGE ¼ 0Þ. The benchmark period
is 1952–1956 (or 1953–1957), which is clearly before the introduc-
tion—let alone the implementation—of the legislation. Thus the
coefficient on CHANGE Y58 measures the impact of the 1956 Act on
changes in innovative output between the benchmark period and the
period 1957–1961 (or 1958–1962); the coefficient on CHANGE Y63
measures the impact of the 1956 Act on changes in innovative output
between the benchmark period and the period 1962–1966 (or 1963–
1967); and so on.
Four key features of the innovations data have directed the choice
of econometric specification. First, the dependent variable takes only
integer values, so a count data model must be used. Second, it is
evident from table 6.1 that there is very considerable overdispersion
24. I also experimented with the capital stock of the average plant, K/N, as a proxy for
setup cost. The results were similar to those reported here. Let me also repeat that
since the model includes industry-specific effects, one need not assume that K/L or K/N
is an accurate measure of setup cost. All that is required is that changes in K/L or K/N
are plausible measures of the change in setup cost.
Price Competition, Innovation, and Market Structure 255
25. This model is of the NB1 form, according to the terminology of Cameron and
Trivedi, that is, the conditional variance of the dependent variable is assumed to be a
multiple of the conditional mean. In addition to a multiplicative individual-specific
fixed effect, the model allows for an individual-specific overdispersion parameter (see
Cameron and Trivedi 1998). As pointed out by Blundell et al. (1995), this model
requires strict exogeneity of the regressors, an assumption which is not implausible in
the present context, where we use industry data and five-year periods. The dynamic
model proposed by Blundell et al. cannot be implemented here, since it requires a long
panel so that presample information can be used to approximate the fixed effects.
256 Chapter 6
CHANGE Y58.26 Because (1) of the time lag between R&D spending
and the commercialization of innovations discussed in section 6.3 and
(2) of the fact that in several R&D-intensive industries price-fixing
agreements were not abandoned until the early 1960s or even later,
any changes in innovative activity between the period 1952–1956
and the period 1957–1962 cannot be attributed to the 1956 Act. If
there is an endogeneity problem with CHANGE that may affect the
regression results, then there should be evidence that the two groups
of industries evolved in different ways during the 1950s.
There is no such evidence from the descriptive statistics in table
6.1. In particular, the mean number of innovations produced in ten
R&D-intensive industries with a subsequent change in competition
regime was 1.8 during 1952–1956 and 3.1 during 1957–1961. The
respective figures for twenty R&D-intensive industries without a
change in regime were 5.4 and 7.6. In other words, there is no evi-
dence that innovative activity changed in different ways in the two
groups of industries during the 1950s: the increase in innovations
between 1952–1956 and 1957–1961 can be attributed to a common
time trend. As we will see below, this is confirmed by the regression
results: the coefficient on CHANGE Y58 is not statistically significant.
We may therefore conclude that any estimated difference between
the two groups in later periods should be due to the 1956 legislation.
Alternatively, if it turns out that there is no difference between the
two groups in later periods, we may conclude that the legislation
had no effect on innovation.
Table 6.2 presents results for the basic sample of thirty R&D-intensive
industries (i.e., industries with average or typical RDS of more than
1% over the period from the mid-1950s to the mid-1970s). The sample
includes ten industries with a change of competition regime. The main
reason for focusing on R&D-intensive industries is the expectation
that any systematic impact of price competition on the production
of innovations may be more difficult to identify or less relevant in
industries where R&D is not a key strategic variable. Also, the theory
of section 6.2 relates specifically to R&D-intensive industries. In any
26. A valid objection to this second type of check is that if CHANGE were indeed
endogenous, then the coefficient on CHANGE Y58 would itself be biased. However,
this check is intended here as a mere confirmation of the picture that emerges from the
descriptive statistics, not as a formal test.
258 Chapter 6
Table 6.2
Regression results for innovation counts in R&D-intensive industries (negative bino-
mial conditional fixed-effects estimation)
Dependent Dependent Dependent
variable: INN1 variable: INN2 variable: INN3
ln SS 0.27 — 0.19 — 0.25 —
(0.16) (0.16) (0.19)
ln DS — 0.17 — 0.11 — 0.23
(0.14) (0.14) (0.17)
ln K/L 0.15 0.07 0.07 0.03 0.28 0.22
(0.26) (0.26) (0.26) (0.26) (0.31) (0.31)
Y58 0.25 0.27 0.33 0.34 0.13 0.13
(0.16) (0.17) (0.17) (0.17) (0.20) (0.20)
Y63 0.19 0.25 0.19 0.23 0.13 0.14
(0.19) (0.19) (0.20) (0.20) (0.23) (0.23)
Y68 0.10 0.20 0.26 0.33 0.20 0.21
(0.26) (0.26) (0.25) (0.25) (0.29) (0.28)
Y73 0.11 0.04 0.14 0.23 0.02 0.01
(0.33) (0.33) (0.32) (0.32) (0.38) (0.37)
CHANGE Y58 0.25 0.26 0.28 0.29 0.21 0.21
(0.36) (0.36) (0.37) (0.37) (0.41) (0.41)
CHANGE Y63 0.26 0.28 0.11 0.12 0.40 0.41
(0.41) (0.40) (0.41) (0.41) (0.46) (0.45)
CHANGE Y68 0.07 0.03 0.07 0.09 0.02 0.02
(0.38) (0.38) (0.39) (0.39) (0.41) (0.40)
CHANGE Y73 0.18 0.11 0.07 0.02 0.02 0.03
(0.39) (0.39) (0.39) (0.39) (0.42) (0.42)
Constant 0.80 0.51 0.16 0.77 0.85 0.57
(1.96) (1.77) (2.01) (1.78) (2.36) (2.09)
Log likelihood 233.0 233.9 241.9 242.3 202.8 202.7
No. of industries 30 30 30 30 30 30
No. of industries 10 10 10 10 10 10
with CHANGE ¼ 1
No. of observations 150 150 150 150 150 150
case, table 6.3 contains results for a larger sample of forty-two in-
dustries, including sixteen with a change of competition regime. This
sample includes, in addition to the R&D-intensive group, a number
of industries with average RDS < 1% but with a relatively large
number of innovations in the SPRU database.
There is no evidence of any overall effect of the intensification of
price competition, following the introduction of the 1956 Act, on
innovations introduced in British industry. The coefficients on the
interaction terms are not statistically significant, even at the 10%
or the 20% level, irrespective of the sample used or the form taken
by the dependent variable (INN1, INN2, or INN3). In particular, the
coefficient on CHANGE Y63 is everywhere negative, but not sta-
tistically significant, while the coefficients on CHANGE Y68 and
CHANGE Y73 are sometimes positive and sometimes negative.
Clearly, the results provide no evidence of any short-run or long-run
effect of price competition on innovation in the present context.
An interesting feature of the results presented in tables 6.2 and 6.3
is the almost general failure of the explanatory variables in these
regressions. Time-invariant industry-specific characteristics seem to
account for much of the cross-industry variation in innovation counts.
As mentioned above, one variable that is almost certainly picked up
by the industry effects is technological opportunity. In addition, a lot
of variation in innovation counts seems to be due to variables diffi-
cult to measure or to observe, including random events.
It is somewhat reassuring that the coefficients on the market size
proxies, ln SS and ln DS, are often statistically significant at the 5%
or the 10% level in regressions using the larger sample (but not in
regressions using the basic sample of R&D-intensive industries). This
suggests that the larger amount of information contained in the
larger sample increases the efficiency of the estimated coefficients
on the market size variables. Note, however, that the magnitude of
these coefficients is fairly low, even in the larger sample. Taken at
face value, these estimates suggest that a 1% increase in market size
causes at most a 0.25% rise in the expected number of innovations.27
27. The rate of growth has sometimes been used instead of market size as a regressor
in empirical models of innovation. To see whether this makes any difference to my
results, I estimated a model including D ln SS or D ln DS among the regressors, defined
for industry i and year t as the change in ln SS or ln DS, respectively, in the five-year
period preceding year t. The coefficients on these variables were everywhere highly
nonsignificant, and the rest of the results did not change much. Note that in order to
use this alternative specification, the first-year observation for each industry had to be
dropped.
260 Chapter 6
Table 6.3
Regression results for innovation counts in the enlarged sample of innovative indus-
tries (negative binomial conditional fixed-effects estimation)
Dependent Dependent Dependent
variable: INN1 variable: INN2 variable: INN3
ln SS 0.27 — 0.19 — 0.25 —
(0.14) (0.14) (0.17)
ln DS — 0.18 — 0.13 — 0.25
(0.12) (0.12) (0.14)
ln K/L 0.19 0.13 0.10 0.06 0.17 0.12
(0.21) (0.21) (0.21) (0.21) (0.25) (0.25)
Y58 0.19 0.21 0.24 0.25 0.07 0.08
(0.14) (0.14) (0.14) (0.14) (0.17) (0.17)
Y63 0.22 0.28 0.22 0.26 0.18 0.18
(0.16) (0.16) (0.17) (0.16) (0.19) (0.19)
Y68 0.12 0.20 0.24 0.30 0.23 0.23
(0.21) (0.20) (0.21) (0.20) (0.24) (0.23)
Y73 0.12 0.01 0.06 0.13 0.03 0.04
(0.27) (0.26) (0.26) (0.25) (0.31) (0.29)
CHANGE Y58 0.11 0.12 0.33 0.34 0.27 0.27
(0.27) (0.27) (0.27) (0.27) (0.30) (0.30)
CHANGE Y63 0.05 0.07 0.03 0.05 0.07 0.07
(0.27) (0.27) (0.29) (0.29) (0.31) (0.31)
CHANGE Y68 0.17 0.19 0.07 0.09 0.28 0.26
(0.28) (0.28) (0.28) (0.28) (0.31) (0.31)
CHANGE Y73 0.12 0.15 0.16 0.17 0.15 0.12
(0.29) (0.29) (0.30) (0.30) (0.32) (0.32)
Constant 0.61 0.42 0.06 0.78 0.48 0.40
(1.67) (1.44) (1.69) (1.45) (1.98) (1.74)
Log likelihood 333.9 334.7 340.3 340.7 294.3 293.9
No. of industries 42 42 42 42 42 42
No. of industries 16 16 16 16 16 16
with CHANGE ¼ 1
No. of observations 210 210 210 210 210 210
28. I also included one industry (industrial engines) with available data for 1958, 1963,
and 1975, but not for 1968.
262 Chapter 6
Table 6.4
The basic samples for the concentration regressions: R&D-intensive industries and
high-R&D industries
No. of No. of No. of No. of
indus- indus- indus- indus-
tries with tries with tries with tries with
CHANGE ¼ 1 CHANGE ¼ 0 CHANGE ¼ 1 CHANGE ¼ 0
and average and average and average and average
Data available for RDS > 1% RDS > 1% RDS > 2% RDS > 2%
All four years 9 18 4 8
1958, 1963, 1968 0 1 0 0
1958, 1963, 1975 1 0 0 0
1963, 1968, 1975 0 22 0 8
1958, 1963 0 3 0 2
1963, 1968 0 6 0 4
Total 10 50 4 22
Table 6.5
Concentration in 1958 and competition regime in R&D-intensive and high-R&D
industries
Mean C5 in 1958
(st. deviation of C5)
Industries with average RDS > 1% ( basic sample)
Industries with CHANGE ¼ 1 ðn ¼ 10Þ 0.704 (0.171)
Industries with CHANGE ¼ 0 ðn ¼ 23Þ 0.683 (0.193)
Industries with average RDS > 1% (more balanced sample)
Industries with CHANGE ¼ 1 ðn ¼ 10Þ 0.704 (0.171)
Industries with CHANGE ¼ 0 ðn ¼ 19Þ 0.686 (0.196)
Industries with average RDS > 2% ( basic sample)
Industries with CHANGE ¼ 1 ðn ¼ 4Þ 0.699 (0.189)
Industries with CHANGE ¼ 0 ðn ¼ 10Þ 0.652 (0.246)
Industries with average RDS > 2% (more balanced sample)
Industries with CHANGE ¼ 1 ðn ¼ 4Þ 0.699 (0.189)
Industries with CHANGE ¼ 0 ðn ¼ 8Þ 0.639 (0.247)
Notes: The figures are based on industries with available data for 1958; n indicates the
number of industries.
Price Competition, Innovation, and Market Structure 263
Table 6.6
Average change in C5, in R&D-intensive and high-R&D industries, 1958–1968 and
1958–1975
DC5 DC5 DC5 DC5
1958–1968 1958–1975 1958–1968 1958–1975
Industries with average RDS > 1%
Industries with 0.093 0.122 0.093 0.118
CHANGE ¼ 1 (0.102) (0.111) (0.102) (0.118)
n¼9 n ¼ 10 n¼9 n¼9
Industries with 0.027 0.034 0.032 0.034
CHANGE ¼ 0 (0.125) (0.122) (0.127) (0.122)
n ¼ 19 n ¼ 18 n ¼ 18 n ¼ 18
Industries with average RDS > 2%
Industries with 0.152 0.111 0.152 0.111
CHANGE ¼ 1 (0.087) (0.096) (0.087) (0.096)
n¼4 n¼4 n¼4 n¼4
Industries with 0.025 0.042 0.025 0.042
CHANGE ¼ 0 (0.092) (0.110) (0.092) (0.110)
n¼8 n¼8 n¼8 n¼8
Notes: The figures in the first column are based on industries with available observa-
tions for both 1958 and 1968. The figures in the second column are based on industries
with available observations for both 1958 and 1975. The figures in the last two columns
are based on industries with available observations for 1958, 1968, and 1975. The fig-
ures in parentheses are standard deviations. n indicates the number of industries.
6.6. Table 6.5 reports descriptive statistics for the concentration ratio,
C5, in 1958 for industries with and industries without a change in
competition regime after 1958. There is no evidence of any significant
difference between the two groups: the mean C5 is 0.704 for the ten
cartelized industries in the sample and 0.683 for the twenty-three
noncartelized industries with available data for that year. On the
other hand, table 6.6 shows that the average increase in C5 after 1958
was much larger for R&D-intensive industries affected by the legis-
lation than for industries not affected. For the former group the
average change between 1958 and 1975 was about twelve percentage
points, while for the latter group it was only 3.4 percentage points.
A comparison of industries that experienced a large increase in C5
between 1958 and 1975 with those that experienced a large decrease
in C5 over the same period is also revealing. The comparison is limited
to the twenty-seven R&D-intensive industries with available data
264 Chapter 6
for both these years. Of the five industries with the largest rise in C5,
three (air and gas compressors; washing machines, electrically oper-
ated; electric lamps) are industries with CHANGE ¼ 1, and only two
(television receiving sets; radio communication equipment) have
CHANGE ¼ 0. However, all five industries with the largest fall in
C5 (cutlery; finished synthetic organic dyestuffs; powered industrial
trucks and industrial tractors; finished detergents; soap) have
CHANGE ¼ 0.
Of course, changes in other potential determinants of concentra-
tion may partly account for these differences. To control for these
other factors, we need to use econometric analysis.
29. They also include an industry with a change of regime but for which C5 is not
available for 1968 although it is available for 1975 (so CHANGE 75 picks up all the
effect of the 1956 Act between 1963 and 1975 in this industry).
Table 6.7
266
Regression results for concentration in industries with average RDS > 1% (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.010 — 0.006 — 0.15 — 0.23 —
(0.015) (0.016) (0.13) (0.14)
ln DS — 0.006 — 0.002 — 0.20 — 0.27
(0.014) (0.014) (0.13) (0.14)
ln K/N 0.073 0.073 — — 0.66 0.66 — —
(0.016) (0.016) (0.15) (0.15)
ln K/L — — 0.062 0.063 — — 0.61 0.62
(0.019) (0.019) (0.19) (0.19)
Y63 0.014 0.015 0.009 0.011 0.15 0.16 0.15 0.16
(0.022) (0.022) (0.022) (0.022) (0.17) (0.17) (0.17) (0.17)
Y68 0.016 0.019 0.008 0.011 0.24 0.28 0.24 0.29
(0.025) (0.025) (0.026) (0.026) (0.20) (0.20) (0.21) (0.22)
Y75 0.024 0.027 0.017 0.020 0.38 0.44 0.41 0.48
(0.029) (0.029) (0.031) (0.031) (0.21) (0.22) (0.24) (0.25)
CHANGE Y63 0.003 0.002 0.002 0.002 0.06 0.05 0.12 0.10
(0.035) (0.035) (0.035) (0.035) (0.24) (0.24) (0.23) (0.23)
CHANGE Y68 0.091 0.092 0.084 0.085 1.10 1.10 1.09 1.09
(0.032) (0.032) (0.033) (0.033) (0.71) (0.71) (0.71) (0.71)
Chapter 6
CHANGE Y75 0.112 0.113 0.104 0.105 0.90 0.91 0.90 0.91
(0.031) (0.031) (0.034) (0.033) (0.32) (0.32) (0.33) (0.32)
Constant 0.786 0.740 0.675 0.634 0.60 1.04 2.08 2.47
267
Table 6.8
268
Regression results for concentration in industries with average RDS > 1% (fixed-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.019 — 0.019 — 0.27 — 0.29 —
(0.023) (0.022) (0.27) (0.27)
ln DS — 0.011 — 0.011 — 0.36 — 0.36
(0.020) (0.020) (0.24) (0.24)
ln K/N 0.030 0.032 — — 0.17 0.14 — —
(0.029) (0.029) (0.29) (0.30)
ln K/L — — 0.018 0.022 — — 0.25 0.21
(0.035) (0.036) (0.41) (0.40)
Y63 0.001 0.002 0.005 0.002 0.07 0.10 0.07 0.10
(0.026) (0.025) (0.025) (0.024) (0.21) (0.21) (0.22) (0.21)
Y68 0.016 0.011 0.025 0.019 0.02 0.09 0.05 0.11
(0.033) (0.032) (0.031) (0.031) (0.29) (0.29) (0.34) (0.33)
Y75 0.019 0.014 0.030 0.022 0.07 0.17 0.13 0.21
(0.040) (0.039) (0.039) (0.038) (0.35) (0.35) (0.44) (0.42)
CHANGE Y63 0.008 0.007 0.009 0.008 0.01 0.02 0.01 0.02
(0.035) (0.035) (0.035) (0.035) (0.29) (0.29) (0.29) (0.29)
CHANGE Y68 0.074 0.076 0.069 0.071 0.97 0.96 0.95 0.95
(0.040) (0.040) (0.040) (0.040) (0.64) (0.63) (0.65) (0.64)
Chapter 6
CHANGE Y75 0.093 0.096 0.087 0.090 0.80 0.82 0.78 0.81
(0.040) (0.040) (0.041) (0.040) (0.43) (0.43) (0.42) (0.42)
R2 0.23 0.22 0.22 0.22 0.15 0.16 0.15 0.16
269
Table 6.9
270
Regression results for concentration in industries with average RDS > 1%: more balanced sample (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.007 — 0.004 — 0.18 — 0.29 —
(0.019) (0.019) (0.18) (0.18)
ln DS — 0.002 — 0.001 — 0.21 — 0.30
(0.017) (0.017) (0.16) (0.16)
ln K/N 0.071 0.071 — — 0.67 0.67 — —
(0.023) (0.023) (0.19) (0.19)
ln K/L — — 0.064 0.065 — — 0.60 0.59
(0.035) (0.035) (0.31) (0.31)
Y63 0.001 0.003 0.003 0.001 0.08 0.10 0.09 0.11
(0.026) (0.026) (0.026) (0.026) (0.21) (0.21) (0.21) (0.21)
Y68 0.009 0.013 0.003 0.006 0.17 0.20 0.17 0.20
(0.030) (0.030) (0.032) (0.032) (0.24) (0.24) (0.27) (0.27)
Y75 0.021 0.025 0.018 0.022 0.51 0.57 0.56 0.61
(0.038) (0.038) (0.044) (0.043) (0.29) (0.30) (0.38) (0.39)
CHANGE Y63 0.013 0.013 0.012 0.012 0.04 0.04 0.03 0.03
(0.037) (0.038) (0.037) (0.037) (0.27) (0.27) (0.25) (0.25)
CHANGE Y68 0.082 0.082 0.076 0.076 0.99 0.99 1.00 0.99
(0.035) (0.035) (0.036) (0.036) (0.74) (0.74) (0.73) (0.73)
Chapter 6
CHANGE Y75 0.106 0.108 0.101 0.103 0.99 1.00 1.02 1.03
(0.036) (0.036) (0.038) (0.038) (0.34) (0.34) (0.35) (0.35)
Constant 0.772 0.718 0.667 0.618 0.86 1.10 2.61 2.72
271
Table 6.10
272
Regression results for concentration in industries with average RDS > 2% (random-effects estimation)
Dependent variable: C5 Dependent variable: logit C5
ln SS 0.006 — 0.011 — 0.16 — 0.39 —
(0.018) (0.020) (0.18) (0.18)
ln DS — 0.013 — 0.017 — 0.27 — 0.47
(0.016) (0.018) (0.16) (0.18)
ln K/N 0.053 0.053 — — 0.81 0.80 — —
(0.029) (0.029) (0.27) (0.27)
ln K/L — — 0.044 0.044 — — 0.76 0.78
(0.038) (0.037) (0.36) (0.36)
Y63 0.024 0.028 0.025 0.029 0.24 0.30 0.37 0.44
(0.032) (0.032) (0.034) (0.033) (0.31) (0.31) (0.33) (0.33)
Y68 0.010 0.017 0.006 0.013 0.31 0.44 0.45 0.60
(0.039) (0.039) (0.042) (0.041) (0.35) (0.36) (0.40) (0.41)
Y75 0.023 0.034 0.021 0.032 0.58 0.77 0.85 1.08
(0.052) (0.053) (0.061) (0.061) (0.40) (0.41) (0.48) (0.51)
CHANGE Y63 0.0001 0.001 0.002 0.002 0.14 0.13 0.26 0.25
(0.055) (0.055) (0.055) (0.056) (0.38) (0.38) (0.39) (0.39)
CHANGE Y68 0.144 0.146 0.134 0.135 2.29 2.29 2.25 2.26
(0.041) (0.041) (0.041) (0.041) (1.45) (1.44) (1.42) (1.41)
Chapter 6
CHANGE Y75 0.111 0.117 0.101 0.106 0.96 1.04 1.04 1.12
(0.056) (0.056) (0.058) (0.057) (0.56) (0.56) (0.57) (0.55)
Constant 0.608 0.546 0.517 0.465 0.82 1.89 3.79 4.61
273
274 Chapter 6
30. The very high coefficients on CHANGE Y68 in these regressions may be partly
due to ‘‘overshooting’’ of the concentration ratio in electrical engineering industries
affected by a series of mergers in the second half of the 1960s, including the 1967–1968
mergers between A.E.I., G.E.C., and English Electric, three of the largest and most
diversified UK electrical engineering firms.
Price Competition, Innovation, and Market Structure 275
31. A third possibility might be that high concentration leads to low (high) innovation,
everything else being equal, so that an increase in the intensity of competition, by
causing concentration to rise, leads to a fall (rise) in innovations. At the same time, the
intensification of price competition could induce firms to innovate more (less) at any
given level of concentration. Again the overall effect could be zero, but the interaction
of the various factors would be more complex than what is implied by the present
analysis.
However, this interpretation does not seem to be consistent with the facts. Because
market structure adjusted only gradually in R&D-intensive industries affected by the
1956 Act, as shown in the concentration regressions of this chapter, some insight on the
interaction of innovation and market structure can be gained by comparing the short-
run effect with the long-run effect of the 1956 Act. The results of tables 6.8 and 6.9,
taken as a whole, provide little evidence that the short-run effect of the Act (i.e., the
effect before the adjustment of market structure) was different from the long-run effect.
276 Chapter 6
7.1 Introduction
period 1951–1958 and the period 1958–1967 for their full sample of
firms, but could not identify any significant difference in the evolu-
tion of profitability between firms in industries affected by the 1956
Act and those in industries not affected. They also found weak evi-
dence of a recovery of the rate of return on capital over the period
1968–1972 in industries affected by the restrictive practices legisla-
tion relative to the control group of noncollusive industries, although
this was not statistically significant. Moreover, a recovery of the rate
of return on capital was also observed for firms in industries with
agreements upheld by the Restrictive Practices Court relative to firms
in the control group. Thus the results provided no evidence of any
effect of the 1956 Act on the profitability of firms. O’Brien et al. (1979)
also examined merger activity over the same period, but again found
no evidence of any significant difference between industries with a
change in competition regime and those without such a change. As
mentioned in chapter 4 of this book, the main methodological limi-
tations of the O’Brien et al. study were the use of a rather small sample
of industries and the fact that the criteria for classifying industries
across groups were somewhat dubious in a few cases.
Despite these limitations, the O’Brien et al. study is one of a few
empirical analyses of the links between competition and profitability
that has explicitly treated concentration as an endogenous variable and
not as a proxy for the degree of market power. The literature on the
effects of economic integration on industry performance has some-
times followed a similar approach, using import penetration or the
degree of tariff protection as measures of competitive pressure and
recognizing the potential effect of changes in these variables on mar-
ket structure as well as on performance. Many of these studies find a
negative effect of import competition on price-cost margins, espe-
cially in concentrated industries, which they typically interpret as
being consistent with the hypothesis that foreign competition reduces
the market power of domestic firms (see, for instance, Jacquemin et al.
1980, deMelo and Urata 1986). There are, however, several important
methodological differences between these studies and the present one.
First, some of the profit measures that are relevant in the context of
the present theory have not been generally used in the existing liter-
ature on the determinants of profitability. Second, economic integra-
tion leads to changes not only in firms’ conduct but also in product
substitutability and market size. Hence it is not easy to draw very
clear implications for the competition-profitability relationship from
Price Competition and Profitability 279
The mechanism that I will test in this chapter has been described in
previous chapters, so I will provide only a brief summary and some
280 Chapter 7
petition regime has no significant effect on plant (or firm) gross or net
profit in the long run.1
1. Also, the effect on industry net profit will be ambiguous, although industry gross
profit should fall because of the fall in the number of firms.
282 Chapter 7
2. See Aghion and Schankerman (2000), where this possibility is discussed along with
many others that are not inconsistent with the predictions of the present theory.
Price Competition and Profitability 283
Two main conclusions can be drawn from the above discussion. First,
in the presence of asymmetries between firms or endogenous sunk
costs, such as advertising or R&D, it is difficult to derive any strong
theoretical results regarding the effect of price competition on prof-
itability in the long run. Second, the basic intuition from the bench-
mark case—that profitability should initially decline following a rise
in the intensity of price competition and then be restored, or par-
tially restored, through a change in market structure—emerges as
the dominant mechanism driving the joint evolution of structure and
performance across classes of industries. The only exception to this
can occur in rather special cases where price competition has a par-
ticularly strong effect on nonprice variables; however, the evidence
from previous chapters suggests that there are very few such cases
in the present context. This has two implications for the empirical
analysis of profitability in this chapter.
The first implication is that the key testable prediction in this
chapter is that profitability should initially decline following an in-
tensification of price competition, and then be restored through a fall
in firm numbers. This will be tested against alternative predictions
derived from two other theories mentioned above: the theory that
cartels deter entry and a strong version of the efficiency differences
approach.
The second implication is that there is probably little to be gained
from performing separate regressions for each class of industries. On
the other hand, there is an important gain from pooling the data,
namely, an increase in the efficiency of the estimation. This is impor-
Price Competition and Profitability 285
tant here for several reasons. First, the effect of price competition on
firm or plant numbers may be less easy to identify than the effect
on concentration, because firm or plant numbers are very sensitive
to the number of small firms or plants, and these may have not been
affected by the 1956 Act to the same extent as larger firms or plants
(see section 7.4). Second, there are few advertising-intensive or R&D-
intensive industries with a change in competition regime in the
present sample. Given that profit data are inherently ‘‘noisy,’’ it is
questionable whether any meaningful results can be derived in sep-
arate regressions for advertising-intensive and R&D-intensive indus-
tries. Third, since one of the key testable predictions in this chapter
is that price competition has no significant long-run effect on profit,
one should use the largest possible sample to give the data the best
possible chance to reveal a nonzero effect. Moreover, another key
aspect of the analysis will be the comparison of short-run and long-
run effects, and again the difficulty of doing this may increase as the
size of the sample used decreases.
In any case, since the theoretical predictions for the effect of price
competition on profit are perhaps less clear for advertising-intensive
and R&D-intensive industries than for exogenous sunk cost indus-
tries, I have also run regressions for the subsample of exogenous
sunk cost industries; the results will be discussed below. The main
purpose of this is to check that the results for the whole sample are
not driven by advertising-intensive or R&D-intensive industries, be-
cause that would not be consistent with the theory.
Before presenting the econometric results, however, I will discuss
some case-study evidence that illustrates very nicely the central
structural mechanism that drives the results of this chapter and of
the book as a whole. In fact, part of the interest of these case studies
derives from the fact that they can confirm that the observed patterns
in the cross-industry data identified in this chapter and in previous
chapters are the result of the mechanism proposed in the theory
rather than of other factors quite different from those suggested by
the theory.
Several of the case studies of the effects of the 1956 Restrictive Trade
Practices Act discussed in Swann et al. (1973) provide evidence that
is consistent with the theory described in section 7.2. Admittedly, the
286 Chapter 7
3. Unfortunately, the Census figures on firm numbers in the transformer industry are
not comparable over the period 1963–1968, because of changes in the definition of
subproducts within the industry.
288 Chapter 7
terpret because the cost of materials seems also to have fallen during
the first half of the 1960s. The clearest indication of a change in
competition regime is therefore the decline in profitability in the
industry in the first half of the 1960s, as documented in Swann et al.
on the basis of data from company reports.
In particular, the largest firm in the industry experienced a con-
tinuous decline in its rate of return to capital, which fell from 18% in
1960 to 5.8% in 1965. For the other firms the evidence is somewhat
mixed, but for all of them profitability clearly declined over 1962–
1964 (i.e., in the first few years after the emergence of price competi-
tion). However, profitability recovered in the mid-1960s. And while
the evidence provided by Swann et al. on the profitability of various
firms in the second half of the 1960s is rather sketchy, the overall
picture is one of sustained moderate profitability for the industry
as a whole. At the same time, prices in early 1969 were at roughly
the same level as in late 1965, although wages and costs of materials
had been rising during that period. Why did the rate of return on
capital recover in the mid-1960s, and how was it then sustained de-
spite increases in input prices that were not fully passed on to selling
prices?
The answer is probably twofold. First, technical progress in the
industry led to increased efficiency during the 1960s and helped to
keep unit costs low despite increases in input prices. Second, the re-
covery and subsequent stability of profits must have been related to
the restructuring of the industry, which was partly realized through
a series of mergers beginning in 1962 and continuing until the late
1960s. Thus, according to highly disaggregated data from the Census
of Production, the number of producers (excluding those with fewer
than twenty-five employees) declined by as much as 50% in all the
principal product lines within the glass container industry between
1958 and 1968, while the five-firm sales concentration ratio for the
industry as a whole increased from about 63.5% in 1958 to 69.5% in
1963, and to 87.3% in 1968, a rise of nearly twenty percentage points
in five years.
The two case studies briefly described above provide support for
the theoretical predictions of the present chapter. Are they typical
examples of the effect of cartel policy on profits and market structure?
This question can be answered only by using econometric analysis.
Prior to that, however, it is worth discussing a counterexample.
290 Chapter 7
Industry Definitions
4. Thus the classification of industries mainly according to supply rather than demand
characteristics is appropriate in the present context because the British cartels typically
operated within industries defined according to their supply characteristics. Also, the
fact that the data cover only UK firms is not a serious problem to the extent that the
impact of the 1956 Act was mostly on competition between UK manufacturers rather
than between manufacturing firms selling in the UK market.
5. Let me clarify this point. The UK Census of Production, like the Censuses of Pro-
duction in many other countries, allocates each plant to an MLH industry (roughly
corresponding to the three-digit level of aggregation) on the basis of the core products
of the plant. MLH industries often are further subdivided into principal products, and
each plant is allocated to a single principal product on the basis of the core product of
the plant. When computing figures for the net value of output, wages and salaries,
firm numbers, and so on, all the products produced in a certain plant are classified to
the industry that the plant is assigned to according to its core product. This means that
the reported net output for the principal product ‘‘coffee,’’ for instance, is not the actual
net output for coffee, but the total net output of all plants whose principal product is
coffee. Sales of coffee by plants classified to other industries are excluded, while sales
of products other than coffee by plants whose principal product is coffee are included.
Obviously, the profit figures, the figures for firm numbers, and so on are all subject to
the same problem.
Note that this occurs only in the data set used in the present chapter, not in any of
the data sets used in previous chapters of this book (except for the two variables that
are defined at the three-digit level in previous chapters, K/N and K/L). For instance, the
sales concentration ratio, the advertising-sales ratio, the value of sales, and so on for
coffee are derived on the basis of the total sales of coffee by all plants in manufacturing
industry.
296 Chapter 7
products, but there are several complications that may arise with
this. I will mention only the most important ones.
One complication is that the 1973 Census of Production, like all
Censuses of Production after 1968, contains no information for prin-
cipal products; all the figures are for entire MLH industries. Hence,
in some cases I had to decide whether to use the figures for a certain
MLH industry for 1954–1973 or the figures for principal products
within that MLH industry for 1954–1968. In practice, the decision to
use principal products was often imposed by the fact that several
MLH industries were ambiguous with respect to the state of compe-
tition, while some of their principal products were not. Whenever
there was a real choice, I usually preferred to use principal products,
since the gain from working with more numerous and more dis-
aggregated industry categories more than offset the loss of informa-
tion for 1973, a year not so crucial for my purposes.6
Second, the definitions of principal products often change across
years, so in several cases the figures are not comparable over the
entire period 1954–1968 (excluding minor noncomparabilities that
can be dealt with by making small adjustments). For instance, there
could be a choice between using an entire MLH as the relevant in-
dustry category, with data for all five years, and using, say, two
principal products within that MLH, with data for 1954–1963 only.
In such cases, I normally chose to use the entire MLH as my industry
definition, unless the MLH as a whole would be classified as ambig-
uous with respect to its state of competition, while some of its prin-
cipal products would not.
A third complication arises when the definition of an MLH in-
dustry changes over time. In some of these cases, it was possible to
construct comparable figures by adjusting the ones published. In
others, I could make use only of the subset of years with comparable
6. Thus, whenever the industry definition that I use is at the four-digit level (i.e., a
principal product) rather than at the three-digit level (i.e., an entire MLH industry), the
1973 observation is not available. Does this lead to any bias in the results regarding the
effect of the 1956 Act by 1973? A bias could be introduced only if MLH industries were
systematically different from principal products in some respect, for instance, in the
degree of homogeneity. But there is no evidence to support such a claim: there are
several fairly heterogeneous industries which are not subdivided into principal prod-
ucts, and hence do have a 1973 observation in my data, just as there are many homo-
geneous industries which are not subdivided into principal products. Moreover, even
if it were the case that industries with an available 1973 observation were different in
some respect from industries without a 1973 observation, this would probably affect
only the estimated time effect for 1973, not the estimated effect of the 1956 Act.
Price Competition and Profitability 297
10. The data set also contains four industries that were affected by the nationalization
of steel in 1967 and had restrictive agreements that were abandoned after 1963. Since
the nationalization of steel made the figures on firm numbers and profitability after
1967 irrelevant, only the 1954–1963 observations for these industries were included
in my data. During this period the collusive agreements were still in place in these
industries, so I chose to classify them as industries with no change in competition re-
gime. Note that the results are barely affected if these industries are dropped from the
sample. I also treated in the same way a fifth industry, telecommunications equipment
(MLH 363), where collusion continued until 1963 and then fully broke down only in
some sections of the industry. Thus I included only the 1954–1963 observations for this
industry in my data set.
11. Recall that since all the empirical models estimated in this book include industry-
specific effects, one need not assume that K/L or K/N is an accurate measure of setup
cost. All that is required is that changes in K/L or K/N are accurate measures of the
change in setup cost, which does not seem implausible.
300 Chapter 7
12. For 1954, the first year in the sample, it was not possible to average relative in-
vestment over two years. Therefore the three-digit industry capital stock was multi-
plied by the ratio of principal product investment to MLH industry investment for
1954.
Price Competition and Profitability 301
in the right direction and (2) that any further refinement of the capi-
tal stock estimates would not affect the results for the variables of
interest, namely, those capturing the competition effect.
Note that an additional important implication of these compari-
sons is that any measurement error in the three-digit industry esti-
mates of K=L and K=N used in previous chapters has not significantly
affected the results for the variables of interest. The results reported
in section 7.5 are those obtained using the adjusted capital stock data.
Of course, one should be aware of the limitations of these data.
These originate not only from my own rough procedure for deriving
estimates at the four-digit industry level, but also from imperfections
in the Census investment figures and the O’Mahony and Oulton
methodology for constructing the three-digit industry capital stock
estimates. These limitations have been discussed in detail in chapter
4, section 4.4. As a result, O’Mahony and Oulton are somewhat cau-
tious about the accuracy of their estimates of the level of capital stock
in different years, and tend to rely more on their estimates of the
change in capital stock across years.
Finally, a remark on the interpretation of the capital-labor ratio in
profit equations may be in order. This variable, or the capital-output
ratio, has often been used in profitability studies to control for the
fact that the endogenous variable, namely, the price-cost margin or
the rate of return on capital, includes the gross return to capital. In
the present study, the capital-labor ratio is seen as a proxy for setup
cost. This is not a real difference, however, since the setup cost is
essentially the cost of installing capital (plant and machinery).
Union power has often been found to have a positive effect on
profitability in empirical studies of the effects of unions, and this re-
sult is consistent with theoretical models of firm-union bargaining. I
therefore include in this chapter a proxy for union power among the
regressors in some specifications. The measure used is union density,
defined as the number of unionized employees over the total number
of employees. Data on this variable for the period examined here are
available at a level of aggregation between the two-digit industry
and the three-digit industry level, and were taken from Bain and
Price (1980).
A potential objection to including a proxy for union power among
the regressors is that this variable could be endogenous. For instance,
the prospect of plant closures following an increase in the intensity
302 Chapter 7
13. For instance, it may be the case that although the monetary benefits from union
membership decrease when the union’s bargaining power in wage negotiations
declines, many workers also think that being union members could reduce their
chances of being laid off.
Price Competition and Profitability 303
The basic sample of industries for this chapter contains 201 indus-
tries and 760 observations. I excluded industries with ambiguous
state of competition in 1958 (or, in a few cases, in the 1960s and early
1970s), as well as industries without at least two available observa-
tions for the core period 1958–1968 (i.e., industries with data avail-
able only for 1954–1958 or 1968–1973). The sample of exogenous
304 Chapter 7
Table 7.1
The samples used in the econometric analysis
No. of No. of
exogenous exogenous
No. of No. of sunk cost sunk cost
indus- indus- indus- indus-
tries with tries with tries with tries with
Data available for CHANGE ¼ 1 CHANGE ¼ 0 CHANGE ¼ 1 CHANGE ¼ 0
All five years 16 30 (1 collusive) 10 15 (1 collusive)
1954, 1958, 1963, 39 49 (1 collusive) 35 32 (1 collusive)
1968
1958, 1963, 1968, 0 2 0 1
1973
1954, 1958, 1963 6 19 (3 collusive) 5 9 (2 collusive)
1958, 1963, 1968 7 1 7 0
1963, 1968, 1973 2 5 2 0
1958, 1963 4 4 (3 collusive) 3 3 (3 collusive)
1963, 1968 3 14 3 9
Total 77 124 65 69
Note: The group of industries without a change in competition regime after 1958
includes three industries where collusion continued throughout 1954–1973 and five
industries where collusion continued for the entire period for which I have data in my
sample (including four industries affected by the nationalization of steel in 1967).
sunk cost industries was derived from the basic sample by dropping
all industries with typical advertising-sales ratio (ADS) or typical
R&D-sales ratio (RDS) higher than 1% over the relevant period; it
contains 134 industries and 502 observations. One difference between
these samples and those used in previous chapters is that I have here
included all industries with available observations for only two of
the three core years 1958, 1963, and 1968—and not just those without
a change of competition regime. Table 7.1 gives details on the struc-
ture of both samples. The full set of industries and information on
key variables can be found in table B6 of appendix B.
Descriptive statistics on initial levels of profit and market structure
measures are presented in table 7.2. In particular, the table reports
means and standard deviations of five different variables in 1954 and
1958, separately for industries with a change in competition regime
after 1958 and industries without such a change. The variables are
the number of firms, NFIRMS; the number of plants, NPLANTS; the
industry gross profit (i.e., the net value of output minus wages and
salaries) divided by the number of plants and deflated by the general
Table 7.2
305
306 Chapter 7
14. It is also interesting in this respect to check the 1954–1958 change of ln NFIRMS
specifically in advertising-intensive and R&D-intensive industries. In each case, there
are only a few industries with CHANGE ¼ 1, so the results should be interpreted with
some caution. Consider first the subsample of R&D-intensive industries with available
data for 1954 and 1958. The average value of ln NFIRMS decreased from 4.55 in 1954 to
4.45 in 1958 in the six industries with CHANGE ¼ 1, while it decreased from 3.47 to
3.36 in the sixteen industries with CHANGE ¼ 0. Clearly, there is almost no difference
between the two groups.
Next, consider the subsample of advertising-intensive industries with available data
for 1954 and 1958. The average value of ln NFIRMS fell from 3.84 to 3.73 in the six
industries with CHANGE ¼ 1, while it fell from 3.80 to 3.67 in the twenty-five indus-
tries with CHANGE ¼ 0. Again, there is very little difference between the two groups.
Price Competition and Profitability 307
On the other hand, the average PCM was roughly constant between
1954 and 1958 in industries with CHANGE ¼ 1 but declined some-
what in industries with CHANGE ¼ 0. This decline was not large, at
least for the full sample: about half a percentage point. Moreover, it
can be argued that if it turns out that the average PCM declined
during 1958–1963 in industries with a change in the intensity of com-
petition relative to the control group, then this must surely be the
effect of the 1956 legislation, since the trend was, if anything, the
opposite before 1958. I will return to these issues when discussing
the regression results below.
Table 7.3 presents statistics on the average change in each of
the five endogenous variables of interest over 1958–1963 and 1963–
1968. (There are considerably fewer observations for 1973 than for
1958–1968, so descriptive statistics for 1958–1973 would be less in-
formative.) In both periods, ln NFIRMS and ln NPLANTS decreased
considerably more, on average, in industries with a change in compe-
tition regime than in industries without such a change. For the period
1958–1968 as a whole, the number of firms in the former group fell by
about 45%, while it fell by only about 20% in the latter group. Also,
the number of plants fell by about 25% in the former group, com-
pared to 8% in the latter. Of course, these comparisons do not control
for changes in other variables; still, the differences between the two
groups are very large indeed.15
With respect to the profit measures, a different picture emerges.
Consider the price-cost margin, which is perhaps the ‘‘cleaner’’ of the
three measures, since it is not directly affected by the changes in
firm and plant numbers. The most interesting figures are those for
the whole sample. In both groups of industries the average PCM in-
creased over 1958–1963 and over 1963–1968. However, during 1958–
1963 the rise was larger for industries with CHANGE ¼ 0 by about
308
Average change in ln NFIRMS, ln NPLANTS, ln PLANTPROFIT, ln FIRMPROFIT, and PCM, 1958–1963 and 1963–1968
D ln NFIRMS D ln NPLANTS D ln PLANTPROFIT D ln FIRMPROFIT DPCM
All industries with CHANGE ¼ 1 ðn ¼ 62Þ
1958–1963 0.18 (0.24) 0.09 (0.23) 0.35 (0.30) 0.44 (0.30) 0.024 (0.033)
1963–1968 0.20 (0.22) 0.15 (0.22) 0.29 (0.25) 0.33 (0.29) 0.012 (0.035)
All industries with CHANGE ¼ 0 ðn ¼ 82Þ
1958–1963 0.09 (0.32) 0.03 (0.34) 0.40 (0.26) 0.47 (0.28) 0.032 (0.033)
1963–1968 0.09 (0.22) 0.05 (0.22) 0.24 (0.31) 0.27 (0.31) 0.004 (0.044)
Exogenous sunk cost industries with CHANGE ¼ 1 ðn ¼ 52Þ
1958–1963 0.19 (0.25) 0.10 (0.24) 0.35 (0.29) 0.44 (0.29) 0.024 (0.030)
1963–1968 0.22 (0.22) 0.16 (0.22) 0.28 (0.26) 0.33 (0.30) 0.011 (0.036)
Exogenous sunk cost industries with CHANGE ¼ 0 ðn ¼ 48Þ
1958–1963 0.12 (0.24) 0.07 (0.26) 0.43 (0.22) 0.48 (0.24) 0.033 (0.022)
1963–1968 0.11 (0.19) 0.08 (0.19) 0.23 (0.29) 0.26 (0.25) 0.012 (0.038)
Notes: The figures are based on industries with available data for 1958, 1963, and 1968. The figures in parentheses are standard deviations.
n denotes the number of industries.
Chapter 7
Price Competition and Profitability 309
one percentage point, while the exact opposite is the case for 1963–
1968. In other words, the average PCM in industries with CHANGE
¼ 1 fell during 1958–1963 by about one percentage point relative to
the control group of industries, but it recovered during 1963–1968.16
In exogenous sunk cost industries, on the other hand, we still have
a relative fall in PCM during 1958–1963 in industries where price
competition intensified, but no obvious recovery during 1963–1968.
Admittedly, these changes are not large, and it is difficult to draw
any conclusions on the basis of these statistics alone. To unravel the
link between changes in market structure and changes in profit-
ability in industries affected by the 1956 Restrictive Trade Practices
Act, I now turn to the econometric analysis.
16. Two of the five industries with the largest increase in PCM between 1958 and 1968
have CHANGE ¼ 1: clay, brick-earth, marl, shale, and chalk; and domestic hollow-
ware. The other three are finished thread; electrical equipment for motor vehicles,
cycles, and aircraft, excluding accumulators (secondary batteries); and powered in-
dustrial trucks and industrial tractors. Similarly, two of the five industries with the
largest decrease in PCM between 1958 and 1968 have CHANGE ¼ 1: cotton waste
yarns; and electrical ware of porcelain, earthenware, or stoneware. The other three are
British wines, alcoholic cider, and perry (fermented pear juice); photographic and
document copying equipment; and margarine.
310 Chapter 7
and similarly for the number of plants. SS is the industry sales reve-
nue deflated by the general producer price index;17 K=L is the capital-
labor ratio; Y54, Y63, Y68, and Y73 are time dummies for 1954, 1963,
1968, and 1973, respectively; and the interaction terms should pick
up any differences after 1958 between industries with a change in
competition regime and industries without such a change. Thus the
coefficient on CHANGE Y63 (CHANGE Y68, CHANGE Y73)
measures the effect of the 1956 Act between 1958 and 1963 (1968,
1973). The benchmark year is 1958, since it is generally accepted that
the Act had little effect on competition before then. The coefficient on
CHANGE Y54 serves as an indirect check of the presumption that
the evolution of market structure during 1954–1958 was not signifi-
cantly different between the two groups of industries. CHANGE is
defined here (and in the profit regressions below) according to the
industry categories used for the endogenous variables. In some
regressions the variable UNION, which denotes union density, is also
included.
The estimated model for each of the three profit measures defined
above is
ln Profit measureit ¼ a i þ g1 ln SSit þ g2 lnðK=xÞit þ g3 Y54 þ g4 Y63 þ g5 Y68
17. As mentioned in section 7.4, the potential endogeneity of prices, especially in profit
regressions, suggests using the general producer price index for all manufacturing as a
deflator of sales revenue (a measure of market size). Using industry-specific price in-
dices as deflators gave results broadly similar to those reported here.
Price Competition and Profitability 311
PCM; ‘‘K/x’’ is either the capital stock of the average plant, K=N, or
the capital-labor ratio, K=L; and the other variables are the same as
above. Again, UNION is sometimes also included as an additional
regressor.18
This specification is very different from those typically used in
‘‘traditional’’ studies of the link between market structure and prof-
itability (such as Cowling and Waterson 1976, or Hart and Morgan
1977), or in the literature on the effects of economic integration on
profitability. In these studies, a measure of market structure is always
included among the regressors. My specification, on the other hand, is
a reduced-form equation derived from a theoretical model in which
market structure and profit are both endogenous. As pointed out in
the introduction to this chapter, the model’s predictions regarding
the effect of a change in the intensity of price competition on profit-
ability depend on allowing market structure to change to restore the
long-run equilibrium. It is therefore important for testing these pre-
dictions that one does not control for changes in market structure
when specifying the profit equation.
Note that a simultaneous-equations approach cannot be used here
because it is difficult to find any variable that affects the number of
firms and does not also influence profitability. Nevertheless, the
reduced-form equations can still provide important insights on the
interaction between market structure and profitability through a
comparison of short-run and long-run effects of competition, as will
be shown later in this chapter.19
18. The denominator of PCM is sales revenue. Some studies (e.g., Hart and Morgan
1977, Conyon and Machin 1991) have used net output as the denominator of PCM on
the grounds that net output, unlike sales revenue, is not influenced by input prices,
duties and subsidies, and the degree of vertical integration within an industry. These
arguments are more important for cross-section studies than for studies using panel
data. In any case, regressions using this alternative definition of PCM gave results
similar to those reported here.
19. Let me also point out again that a dynamic panel data model cannot be used in the
present context because of data limitations. Since the years in my panel are separated
by periods of four to five years, however, it is not clear that there should be any sig-
nificant effect of lagged values of the endogenous variables because of adjustment lags
or for other reasons. Also, some of the other arguments advanced in the context of the
concentration regressions of chapter 4 may be of relevance here as well, namely, that
the econometric specification used allows the competition effect to operate with a lag,
that setup costs are measured somewhat imprecisely anyway, and that there is no
evidence of significant serial correlation in the residuals of the models estimated in
this chapter.
312 Chapter 7
Estimation Results
The model has been estimated for the whole sample as well as for the
sample of exogenous sunk cost industries and the results are presented
in tables 7.4–7.6. All the results are for a fixed-effects specification.20
The reported standard errors are heteroskedasticity-consistent, ad-
justed for small sample bias following MacKinnon and White (1985).
As in previous regressions using fixed-effects estimation, two differ-
ent R 2 ’s are reported: the first does not include the fixed industry
effects, while the second does.21
Table 7.4 contains regression results for ln NFIRMS and
ln NPLANTS. Note that in the regressions using the whole sample,
interaction variables are used to control for possible differences be-
tween exogenous sunk cost, advertising-intensive, and R&D-intensive
industries with respect to the effect of market size on the number of
firms or plants. In particular, AD2 ln SS ðRD2 ln SSÞ is equal to
ln SS for industries with typical or average ADS (RDS) higher than
2% over the relevant period and to 0 for industries with ADS (RDS)
lower than 2%.22
The results in table 7.4 suggest that the 1956 Act had a strong
and statistically significant negative effect on the number of firms
in the long run. This effect was only partly realized by 1963, and it
was mostly realized by 1968. The magnitude of the coefficient on
CHANGE Y73 implies that the intensification of price competition
following the 1956 Act reduced the number of firms by about 12%
in the average industry between 1958 and 1973. For exogenous sunk
20. The Hausman test always rejects the random-effects model. In any case, the results
from this model with respect to the competition effect are similar to those obtained
using fixed-effects estimation.
21. The reported results are for models with standard i.i.d. residuals. I also estimated a
model with an AR1 error structure, and the results were very similar to those reported
here. The estimated coefficient of serial correlation in these regressions was very low
(and often even negative): around 0.05 in market structure regressions, and between
0.05 and 0.20 in profit regressions.
22. I also experimented with alternative interaction variables—using 1% instead of 2%
as the cutoff point—but these were not statistically significant. In preliminary regres-
sions, I also included such interaction variables for the competition effect, but they
were not statistically significant, either individually or jointly. Note that whether
an industry’s typical advertising-sales ratio or R&D-sales ratio over a period of ten or
twenty years is higher or lower than 2% is largely determined by exogenous charac-
teristics such as advertising effectiveness and technological opportunity, so the inter-
action variables are not endogenous.
Price Competition and Profitability 313
23. It is not surprising that differences across classes of industries with respect to
the market size effect on firm numbers are not significant. Numerous small firms in
advertising-intensive and R&D-intensive industries spend little or nothing on advertis-
ing or R&D, and may produce secondary industry products. The endogenous sunk cost
models of chapters 5 and 6 (or of Sutton 1991, 1998) are not relevant for these firms.
Table 7.4
314
Regression results for ln NFIRMS and ln NPLANTS (fixed-effects estimation)
Dependent variable: ln FIRMS Dependent variable: ln PLANTS
All industries Exogenous sunk cost industries All industries Exogenous sunk cost industries
ln SS 0.582 0.581 0.518 0.515 0.658 0.658 0.605 0.604
(0.041) (0.041) (0.041) (0.040) (0.031) (0.031) (0.031) (0.031)
ln K=L 0.065 0.064 0.108 0.110 0.125 0.124 0.138 0.138
(0.047) (0.047) (0.055) (0.055) (0.041) (0.041) (0.048) (0.048)
UNION — 0.346 — 0.514 — 0.209 — 0.194
(0.293) (0.381) (0.226) (0.286)
Y54 0.113 0.119 0.080 0.086 0.100 0.103 0.063 0.065
(0.024) (0.025) (0.030) (0.030) (0.024) (0.024) (0.028) (0.029)
Y63 0.182 0.177 0.191 0.184 0.102 0.099 0.114 0.111
(0.022) (0.022) (0.026) (0.026) (0.021) (0.022) (0.025) (0.025)
Y68 0.351 0.340 0.322 0.310 0.234 0.228 0.226 0.222
(0.034) (0.034) (0.041) (0.040) (0.032) (0.032) (0.039) (0.038)
Y73 0.450 0.408 0.401 0.355 0.381 0.356 0.319 0.302
(0.051) (0.059) (0.061) (0.067) (0.045) (0.052) (0.053) (0.056)
CHANGE Y54 0.004 0.005 0.028 0.030 0.054 0.054 0.028 0.027
(0.040) (0.040) (0.045) (0.045) (0.035) (0.035) (0.041) (0.041)
CHANGE Y63 0.037 0.035 0.021 0.017 0.025 0.024 0.014 0.012
Chapter 7
(0.031) (0.031) (0.035) (0.035) (0.029) (0.029) (0.034) (0.033)
CHANGE Y68 0.095 0.086 0.110 0.090 0.063 0.057 0.062 0.054
(0.042) (0.043) (0.048) (0.052) (0.035) (0.036) (0.040) (0.043)
CHANGE Y73 0.133 0.119 0.190 0.155 0.080 0.071 0.128 0.115
315
Table 7.5
316
Regression results for ln PLANTPROFIT and ln FIRMPROFIT (fixed-effects estimation)
Dependent variable: ln FIRMPROFIT Dependent variable: ln PLANTPROFIT
All industries Exogenous sunk cost industries All industries Exogenous sunk cost industries
ln SS 0.441 0.440 0.455 0.453 0.363 0.362 0.365 0.362
(0.058) (0.058) (0.059) (0.058) (0.054) (0.054) (0.051) (0.050)
ln K=L 0.146 0.147 0.154 0.152 0.200 0.202 0.182 0.179
(0.055) (0.055) (0.065) (0.065) (0.052) (0.052) (0.059) (0.060)
UNION — 0.532 — 0.446 — 0.638 — 0.767
(0.389) (0.421) (0.316) (0.361)
Y54 0.078 0.070 0.018 0.012 0.068 0.058 0.003 0.007
(0.036) (0.036) (0.039) (0.039) (0.035) (0.035) (0.038) (0.037)
Y63 0.347 0.354 0.379 0.385 0.268 0.277 0.302 0.314
(0.031) (0.031) (0.036) (0.036) (0.031) (0.030) (0.036) (0.036)
Y68 0.524 0.542 0.575 0.585 0.411 0.431 0.481 0.498
(0.046) (0.046) (0.052) (0.051) (0.044) (0.044) (0.050) (0.049)
Y73 0.616 0.680 0.694 0.734 0.566 0.641 0.616 0.685
(0.067) (0.070) (0.075) (0.077) (0.059) (0.066) (0.069) (0.074)
CHANGE Y54 0.043 0.042 0.085 0.083 0.017 0.018 0.026 0.024
(0.057) (0.057) (0.062) (0.062) (0.054) (0.054) (0.060) (0.059)
CHANGE Y63 0.014 0.011 0.031 0.028 0.026 0.022 0.038 0.033
Chapter 7
(0.046) (0.046) (0.050) (0.050) (0.045) (0.045) (0.050) (0.050)
CHANGE Y68 0.069 0.082 0.028 0.045 0.035 0.052 0.020 0.009
(0.055) (0.057) (0.061) (0.065) (0.050) (0.051) (0.056) (0.059)
CHANGE Y73 0.092 0.114 0.115 0.145 0.012 0.040 0.035 0.088
317
318 Chapter 7
Table 7.6 reports the results. The first thing to note is that the
coefficients on CHANGE Y68 and CHANGE Y73 are small and
nowhere statistically significant, even at the 10% level. The failure to
detect any long-run effect of price competition on the price-cost
margin is consistent with the theory developed in this book. In par-
ticular, it justifies the Selten-Sutton emphasis on the effect of firm
conduct on market structure rather than on profits. Of the other
variables, ln K=L and ln K=N have a positive effect on PCM, while
union density has a negative effect, as expected. The time dummies
may again be picking up some of the effect of scale economies on the
price-cost margin, given that ln K=L and ln K=N are imperfect proxies
for setup cost. These results are hardly affected if the sales variable,
which is nowhere statistically significant, is dropped.
It is also very interesting to compare the short-run and the long-
run impact of competition. This may be the most decisive test of the
present theory, which predicts a particular link between the evolu-
tion of market structure and the evolution of profitability in the short
run and in the long run. Table 7.4 shows that the effect of the 1956
Act on market structure was modest over the period 1958–1963;
on the other hand, it was between 1963 and 1968 that most of the
restructuring of previously cartelized industries occurred. Moreover,
the overall picture from table 7.6 (despite small differences across
regressions) is that price-cost margins declined, on average, between
1958 and 1963 in these industries, before recovering mostly during
1963–1968. Note that the coefficient on CHANGE Y63 is everywhere
negative and typically statistically significant at the 5% or the 10%
level, while the coefficients on CHANGE Y68 and CHANGE Y73
are sometimes positive, sometimes negative, and nowhere statisti-
cally significant.
This is precisely the sort of link the theory predicts between the
evolution of market structure and the evolution of profitability in
previously collusive industries: a moderate effect on the number of
firms by 1963, at which date several industries were in short-run
disequilibrium with reduced margins (including industries where
the adjustment of concentration to its long-run value was being
delayed by a slow rate of depreciation of the capital stock); then a
significant negative effect on firm numbers between 1963 and 1968,
leading to a rise in price-cost margins in these industries.
The coefficient on CHANGE Y63 in regressions with PCM, al-
though typically significant at the 5% or 10% level, is nevertheless
Table 7.6
320
Regression results for PCM (fixed-effects estimation)
Dependent variable: PCM
All industries Exogenous sunk cost industries
ln SS 0.008 0.006 0.008 0.005 0.004 0.002 0.003 0.001
(0.007) (0.006) (0.006) (0.006) (0.007) (0.007) (0.006) (0.006)
ln K=L 0.020 — 0.021 — 0.014 — 0.014 —
(0.007) (0.007) (0.008) (0.008)
ln K=N — 0.012 — 0.013 — 0.007 — 0.007
(0.006) (0.006) (0.007) (0.007)
UNION — — 0.177 0.176 — — 0.186 0.188
(0.042) (0.041) (0.051) (0.051)
Y54 0.006 0.005 0.008 0.008 0.010 0.009 0.012 0.011
(0.004) (0.004) (0.004) (0.004) (0.005) (0.005) (0.005) (0.005)
Y63 0.029 0.030 0.031 0.033 0.030 0.032 0.033 0.034
(0.004) (0.004) (0.004) (0.004) (0.005) (0.005) (0.005) (0.005)
Y68 0.026 0.030 0.032 0.035 0.036 0.040 0.040 0.043
(0.006) (0.006) (0.005) (0.005) (0.006) (0.006) (0.006) (0.006)
Y73 0.021 0.030 0.042 0.050 0.040 0.046 0.057 0.062
(0.009) (0.009) (0.009) (0.009) (0.011) (0.010) (0.010) (0.009)
CHANGE Y54 0.003 0.003 0.003 0.003 0.005 0.005 0.005 0.005
Chapter 7
(0.007) (0.007) (0.006) (0.006) (0.007) (0.007) (0.007) (0.007)
CHANGE Y63 0.012 0.012 0.011 0.011 0.011 0.011 0.009 0.009
(0.006) (0.006) (0.006) (0.006) (0.007) (0.007) (0.007) (0.007)
CHANGE Y68 0.004 0.004 0.001 0.001 0.010 0.010 0.003 0.003
321
322 Chapter 7
not large: about one percentage point. This is consistent with the
evidence from the descriptive statistics of table 7.3. It has to be borne
in mind that only a subset of the previously cartelized industries
were in short-run disequilibrium in 1963. In several industries com-
petition had not yet emerged, and there must have been several
others where competition had emerged and much of the adjustment
of market structure had already taken place. Thus the magnitude of
the coefficient on CHANGE Y63 should not be taken as a measure of
the fall in the price-cost margin following a change of competition
regime and prior to any adjustment of market structure; this is likely
to be much larger than one percentage point.
Several additional remarks are in order. First, the econometric
analysis confirms that the potential endogeneity of CHANGE does
not seem to be a source of significant bias in the results. Recall that
an indirect check of this claim is to compare the evolution of market
structure and profitability between industries with CHANGE ¼ 1
and industries with CHANGE ¼ 0 before 1958. I have already empha-
sized that the descriptive statistics provide no evidence that the dif-
ferences between the two groups after 1958 could be attributed to a
preexisting differential trend. This is confirmed by the econometric
analysis. Thus the coefficient on CHANGE Y54 is nowhere statisti-
cally significant, even at the 20% level, suggesting that there was no
difference in the evolution of market structure and profitability be-
tween the two groups of industries before 1958 (the benchmark year).
Moreover, the coefficient on CHANGE Y63 usually has the same sign
as the coefficient on CHANGE Y54, which implies that even if a
differential trend between the two groups of industries had existed
before 1958 for some variables and is imprecisely measured (because
the 1954–1958 period is too short, say), this trend was, if anything,
reversed during 1958–1963.
Second, a comparison of the R 2 ’s in table 7.4 with those in table 7.6
reveals an interesting feature of the regressions presented in this
chapter. In both these tables two R 2 ’s are reported and, as pointed
out previously, the difference between the two is a measure of the
explanatory power of the industry effects. It can be seen that when
the R 2 is defined to exclude fixed industry effects, it is very much
higher in regressions with ln NFIRMS and ln NPLANTS than in re-
gressions with PCM. On the other hand, when the R 2 is defined to
include fixed industry effects, it is only slightly higher in regressions
Price Competition and Profitability 323
24. Another interesting implication of the results presented in this chapter is that they
confirm that explicit collusion was generally not replaced by tacit collusion in the long
run in the large majority of the previously collusive industries. Recall that this is also
what the case-study evidence suggests, as pointed out in chapter 2. If explicit collusion
had been replaced by tacit collusion, then the profitability of industries affected by the
1956 Act would have increased in the long run relative to the profitability of industries
not affected by the legislation, given that the fall in firm numbers was more pro-
nounced in the former group than in the latter. But this is not what we observe.
Price Competition and Profitability 325
Wider Issues
the distinction between ‘‘minimum’’ and ‘‘common’’ prices carries too much
weight—especially in light of the fact that many agreements were redrafted
prior to registration. I therefore will often simply use a term such as ‘‘pricing
agreement’’ or ‘‘price-fixing agreement’’ or an equivalent expression in what
follows. On the other hand, I will always specifically mention any important
nonprice restrictions, such as market sharing, collective exclusive dealing,
aggregated rebates, and restrictions relating to advertising or R&D.
Whenever a particular agreement did not provide for price-fixing or mar-
ket sharing, I will elaborate on the types of restrictions that were in force. I
will not, however, discuss cases of individual resale price maintenance by
firms in an industry, either before or after 1956. These do not constitute
agreements under the 1956 Act; more important, as I explained in chapter 3,
there is no evidence that resale price maintenance had a significant effect on
competition between manufacturers in the absence of any other restrictions.
The question of whether a particular agreement covered the export market
in addition to the home market is often not an easy one, so it will not be
given much attention here, except for very few cases where the relevant in-
formation is both available and important for classifying an industry as col-
lusive, competitive, or ambiguous in the 1950s. Another difficult question
concerns the participation of British firms and associations in international
cartels. Since this issue is not directly relevant to the present study, it will be,
for the most part, left aside. I will report only significant cases of inter-
national cartels that are mentioned in the Register of Restrictive Trading
Agreements or the reports of the Monopolies and Restrictive Practices Com-
mission. On the other hand, I will report all the available evidence on infor-
mation agreements—although the lack of systematic data on information
agreements poses some serious problems in this respect. Thus my review of
information agreements, based on evidence contained in the register or from
case studies, will inevitably be incomplete.
The available information on significant restrictions other than pricing
restrictions is also less than complete. The main reason is that several agree-
ments were modified before registration, and this sometimes implied remov-
ing provisions regarding market sharing, collective exclusive dealing, or
aggregated rebates. This difficulty should not be overemphasized, however,
since information on such schemes is provided in several data sources other
than the register, including the 1955 Monopolies and Restrictive Practices
Commission report on ‘‘collective discrimination.’’ This report covered the
large majority of the most important instances of collective exclusive dealing,
aggregated rebates, barriers to entry into distribution, and collective resale
price maintenance—although it did not usually specify the exact restrictions
that applied to each industry listed, provided little information on coverage,
and sometimes used rather broad industry definitions.1
1. There are four such broad industry definitions in the report: ‘‘certain chemicals,’’
‘‘certain electrical products,’’ ‘‘groceries,’’ and ‘‘hardware’’; the last two should be taken
to mean ‘‘certain groceries’’ and ‘‘certain hardware products.’’
Survey of Collusive Agreements in British Industries 335
The following abbreviations will be used for some of the data sources.
Rxxx denotes the number (xxx) of an agreement in the Register of Restrictive
Trading Agreements. BT denotes the Board of Trade annual reports on the
operation of the 1948 Monopolies and Restrictive Practices Act, covering the
years 1949 to 1956. PEP denotes the Political and Economic Planning survey
of industrial trade associations, including the unpublished background ma-
terial for this survey. MRPC stands for Monopolies and Restrictive Practices
Commission, MC stands for Monopolies Commission, and MMC stands for
Monopolies and Mergers Commission (all of which are successive names
for the same public body). ‘‘Lloyds’ report’’ refers to the 1949 report of the
Lloyds’ Committee on resale price maintenance (Board of Trade 1949), and
CD refers to the 1955 MRPC report on collective discrimination. Finally, the
reports of the Restrictive Practices Court, which have been published as
Reports on Restrictive Practices Cases by the Incorporated Council of Law
Reporting, are conventionally cited as LR [volume no.] RP [page no.].
For each product, I will start with the information contained in the Register
of Restrictive Trading Agreements and then report any relevant additional
information from other sources. To avoid repetition, references to other data
sources will be made only when these sources contain additional informa-
tion relative to the register. This is typically the case with the reports of the
Monopolies and Restrictive Practices Commission, the reports of the Re-
strictive Practices Court, the 1955 MRPC report on collective discrimination,
and the case studies contained in Swann et al. (1973). It is less often the case
with the lists of collusive or allegedly collusive industries in PEP and BT.
Hence references to PEP or BT will be made below only (1) for industries
which did not register any significant restrictive agreements but are listed as
collusive or allegedly collusive in these sources, and (2) for industries whose
registered agreements do not seem significant and the detailed evidence
contained in PEP confirms that they should be treated as competitive. The BT
lists distinguish several types of restrictive practices, including price-fixing,
market sharing, and collective discrimination. Unless otherwise stated, ‘‘col-
lusion’’ should be taken to mean price-fixing when a reference is made to the
BT lists below.
I will also indicate which agreements were referred to the Restrictive
Practices Court. In particular, agreements referred to the Court will be
marked with an * next to their number on the register (i.e., Rxxx*). For those
that were struck down by the Court without being defended by the parties,
no other information will be given.2 For those that were contested by the
firms involved, I will indicate what the outcome was in each case. Brief
descriptions of all the agreements that were referred to the Court since the
introduction of the 1956 Act and summary accounts of Court proceedings for
2. I will not distinguish between agreements abandoned after being referred to the
Court (which was the usual case) and agreements referred to the Court after being
abandoned (which was a precautionary step sometimes taken by the Registrar in order
to reduce the likelihood of any future collusion by the firms involved). This distinction
is usually difficult to make on the basis of the available data.
336 Appendix A
Slate and slate products. A pricing agreement of the North Wales Slate
Quarries Association (R400) was abandoned in 1959. Wales accounted for
about 80% of the total UK production of slate and slate products in the 1950s.
Hence there was probably a change of competition regime in this industry.
Sand and gravel. About thirty regional agreements in sand and gravel
were registered, then formally abandoned between 1959 and 1961—although
some of them were replaced by information agreements. They all comprised
agreed prices, and some also provided for market sharing or the allocation
of work. There were also agreements between associations to observe each
other’s arrangements in their respective areas. The agreements probably
covered more than half of total industry sales; this is also the view expressed
in PEP. Three of the agreements (R595*, R943*, R1712*) were by Scottish
associations that had previously been the subject of an official inquiry—see
MRPC, Report on the Supply of Sand and Gravel in Central Scotland (London:
H.M.S.O., 1956). An official investigation into the sand and gravel industry
carried out in the mid-1970s revealed several dozen regional or local restric-
tive agreements which had been in operation since the early 1970s or even
earlier. These agreements were placed on the register (and most were re-
ferred to the Court), and were abandoned in the late 1970s. There is some
uncertainty about their coverage and effectiveness, however. In summary,
the state of competition in the industry during the 1960s and the 1970s
should probably be regarded as ambiguous.
Clay. Three national agreements covered all types of clay except fireclay.
The China Clay Association recommended price changes and required mem-
bers to report their individual prices and any intention to change them to
the association (R128). The China Stone Association fixed prices and quotas
(R190). And the Ball Clay Producers Federation’s agreement contained min-
338 Appendix A
imum list prices, although it did not regulate trade discounts (R377). All
three agreements were abandoned in 1959. In addition, various associations
of firebrick producers (listed under MLH 461 below) fixed the price of fire-
clay until 1959–1961. Overall, there was a change of competition regime in
the clay industry.
Iron ore. The case of iron ore is special in several ways. First, imports were
far greater than domestic production and they were centrally purchased.
Second, the Iron and Steel Board (the regulatory body for the steel industry)
had implemented a scheme that may have resulted in price discrimination
against users of domestic iron ore. Third, most of the domestic iron ore was
mined by steel firms, which were also the principal users. See D. Burn, The
Steel Industry 1939–1959 (Cambridge: Cambridge University Press, 1961) for
details. In any case, there is no evidence of any collusive agreements between
iron ore producers, and there was no change in competitive conditions be-
tween the mid-1950s and the mid-1960s. In 1967 the iron ore industry was
radically transformed as a result of the nationalization of steel.
Salt. The salt industry is mentioned as collusive in PEP, and it was also
listed in CD; however, no agreement was registered in the 1950s. An inves-
tigation of the industry by the MMC in the 1980s revealed an unregistered
pricing agreement, which had been in operation for several decades. See
MMC, White Salt (London: H.M.S.O., 1986). Thus the salt industry was col-
lusive throughout the time period examined in this book.
ommended prices (R175, R176*, R776*, R777, R778, R779, R780, R783*, R969,
R1247, R1599*, R1608, R1652*, R2276, R2429, R2443). There were also five
important national agreements. Three of them, by the National Association
of British and Irish Millers, covered most types of flour and associated
products, and comprised recommended prices and a scheme of aggregated
rebates for large buyers (R782*, R784*, R844). The fourth, by the Association
of Millers of Proprietary Brown Flour, required members to report their in-
dividual prices and any intention to change them to the association (R843).
The fifth, by the Millers Mutual Association, whose members included all the
important producers of flour, comprised a system of quotas (R2154). All the
national and regional agreements were abandoned in 1958–1959. Hence
there was a change of competition regime in the flour industry. An MMC
inquiry found competition in the industry to be effective in the 1970s—see
MMC, Flour and Bread (London: H.M.S.O., 1977).
Bread. The price of bread was under government control until 1956. After
1956, it became subject to various agreements by national, regional, and local
associations. The three most important agreements were referred to the Re-
strictive Practices Court and were unsuccessfully contested by the parties
(see LR 1 RP 387, LR 1 RP 347). In particular, R962* was a pricing agreement
by the Federation of Wholesale and Multiple Bakers, whose members were
‘‘plant bakers’’ in England and Wales. R252*, R981*, and R980* were price
agreements by the Wholesale and Retail Bakers of Scotland Association, the
Scottish Association of Master Bakers, and a Joint Costing Committee of the
two associations, respectively. All these agreements were abandoned in
1959–1960. Regional agreements in the industry (R985, R1048, R1311, R2462,
R2469, R2470, R2471, R2478, R2480, R2580, R2599, R2601) were also termi-
nated at that time, as were agreements by associations of ‘‘master bakers’’
(R1542, R1566, R2275) and numerous local collusive arrangements.
Swann et al. (1973) describe the gradual emergence of competition in the
bread industry, despite parallel pricing and some information agreements
about discounts—for instance, R3061*, which ended in 1965. Further infor-
mation is contained in MMC, Flour and Bread (London: H.M.S.O., 1977). The
MMC discovered a large number of agreements between leading firms that
had been in operation since the late 1960s or early 1970s. Most of them
involved the exchange of information about discounts to particular buyers
or about intended increases in discounts. Some arrangements were national,
340 Appendix A
but most were local. All of them were placed on the register (and some were
referred to the Court) and were abandoned in the mid-1970s. See the report
of the Director General of Fair Trading for 1977 for details. However, it is
very doubtful that these arrangements had any significant effect on compe-
tition. The view of the MMC was that they were an attempt, largely unsuc-
cessful, to resist pressure from retailers for progressively larger discounts.
In fact, discounts kept rising between 1965 and 1975, which suggests that
competition was largely effective. Hence the bread industry experienced
a change of competition regime as a result of the 1956 Act.
Flour confectionery. This industry was not subject to any national agree-
ments. Some of the regional or local pricing arrangements among bakers
covered flour confectionery as well, but most did not. These agreements
probably affected more than 10% but less than 50% of total industry sales.
Bacon. The British Bacon Agents Association recommended prices for Brit-
ish bacon (R2750). Its members were agents appointed by curers or curers
who acted as their own agents. The agreement ended in 1961 and was re-
placed by an information agreement, which lasted until the early 1970s.
However, about a third of the members withdrew from the association in
1962–1963. A pricing agreement by the Association of Scottish Bacon Curers
and the Scottish Provision Trade Association was abandoned in 1960
(R1185*). Finally, the register also contains a series of annual contracts for the
purchase of pigs between farmers’ representatives and the British Bacon
Curers Federation or smaller associations of curers (R2301, R2305, R2286).
These contracts, which specified prices and terms of purchase, had been a
feature of the industry ever since the end of government control of the dis-
tribution of fatstock, and continued until the early 1970s. It is unlikely that
they had any significant effect on competition between producers of bacon.
In conclusion, there was a change of competition regime in this industry.
Survey of Collusive Agreements in British Industries 341
Butter. The register contains two national pricing agreements: R814 by the
English Butter Conference and R1047 by the Butter Makers and Packers As-
sociation. Both ended in 1959–1960. Hence there was a change of competi-
tion regime in the butter industry.
formation agreement. The fraction of the cheese industry affected by the 1956
Act was probably around 40%.
Prices, market shares, and profit margins in the sugar industry were under
government regulation throughout the period examined in this book. Details
can be found in Hart et al. (1973) and in MMC, S & W Berisford Limited and
British Sugar Corporation Limited: A Report on the Proposed Merger (London:
H.M.S.O., 1981). All the registered agreements relating to sugar were among
distributors, except one, which was a pricing scheme implemented by the
government and the three main manufacturers in 1974 (R3833). It replaced
previous schemes in the industry and was discontinued in 1976.
Animal feeding stuffs, other than dog and cat foods. No important agree-
ments were registered in the 1950s. (An agreement by the Association of Fish
Meal Manufacturers, R1635, related only to delivery charges; besides, fish
meal is a minor product of the industry). However, the industry was listed
in BT as allegedly collusive. In addition, an official investigation carried
out in the 1970s revealed that a collusive agreement between leading pro-
ducers had been in operation since 1970 (R4556*, R4673*). This was aban-
doned in 1978. All in all, the effect of the 1956 Act on this industry was
ambiguous.
344 Appendix A
Dog and cat foods. The National Association of Dog Biscuit Manufacturers
fixed the price of biscuit powder or dust, biscuit meals, and dog meals (R96).
The agreement dealt only with residuals and low-priced dog and poultry
food, not with the manufacturers’ proprietary brands. Since the whole of the
product category ‘‘dog biscuits’’ (including proprietary brands) accounted
for only about 10–20% of total sales of dog and cat foods during most of the
relevant period, the industry can be seen as essentially competitive.
Coffee, and coffee and chicory extracts and essences. No evidence of any
restrictive agreements.
Whisky. Domestic and export sales of whiskey were officially rationed until
1954. In fact, rationing in the home market continued until the late 1950s (see
Hart et al. 1973). Moreover, the firms agreed on maximum prices for whiskey,
according to CD, PEP, and BT, although no such agreement appears in the
register. The effect of rationing and of maximum prices on competition is
difficult to determine in general, so the state of competition in this industry
in the 1950s should probably be seen as ambiguous.
The principal sources of information on the tobacco industry are PEP (which,
in this case, reports evidence given by a tobacco firm), the Lloyds’ report,
and especially MC, Report on the Supply of Cigarettes and Tobacco and of Ciga-
rette and Tobacco Machinery (London: H.M.S.O., 1961). According to these
sources, the most important factor restricting competition in the industry
during the early 1950s was the government-controlled allocation of raw
materials among producers. This ended in 1954, and was followed by a large
expansion in the market share of the two leading firms at the expense of
smaller firms. An agreement by the Northern Tobacco Manufacturers Asso-
ciation to set minimum prices of cheaper types of tobacco sold in Scotland
and northern England also ended at that time; in any case, the largest firm in
the industry had never been a party to that arrangement. After 1954, there
were no horizontal agreements on prices, although the industry practiced
collective resale price maintenance through the Tobacco Trade Association
until 1956 (see also CD and BT). There were also occasional discussions
about distributors’ margins, but no explicit agreement. The register does not
contain any important restrictive agreements among tobacco manufacturers.
This evidence suggests (1) that the state of competition in the industry was
ambiguous until the mid-1950s because of government control, and (2) that it
would probably be safe to classify the industry as one without a change in
competition regime between the late 1950s and the 1970s.
Survey of Collusive Agreements in British Industries 347
Coke. The register contains two national and several regional agreements
on coke. The price of blast-furnace coke was regulated by an agreement
between the British Coking Industry Association, the British Iron and Steel
Federation, and the Council of Iron Producers (R649*). The British Coking
Industry Association also recommended prices of hard coke and had mar-
keting schemes for various regions (R650*). Both agreements ended in 1964–
1965. Further information on the hard coke industry is provided in PEP,
in which it is pointed out that the retail prices of hard coke were regulated
by the government, while the producer prices fixed by the British Coking
Industry Association were implemented as part of a government-approved
scheme. What is important for our present purposes is that during 1954–
1963, there was no change in the state of competition in the coke industry
(hard coke and blast-furnace coke). After the mid-1960s, conditions changed
dramatically as a result of the nationalization of the steel industry, since the
steel firms produced about half of the total output of coke. The register also
contains a series of price agreements for gas coke, which were abandoned in
the early 1960s. However, gas coke is not part of MLH 261.
the MC noted that price competition intensified after 1960, but it attributed
this to the entry of new firms. All in all, the effect of the 1956 Act on com-
petition in the mineral oil refining industry was ambiguous.
Products of coal tar distillation. Almost all the products in this industry
were subject to collusion. The Association of Tar Distillers recommended
prices for road tar and bitumen tar mixtures, coal tar fuels, and certain cre-
osote oils (R1033). The agreement was canceled in 1960 and replaced by an
information agreement, which ended in 1965. Refined tar was also subject to
regional agreements until 1960 (R1417, R1458). The Anthracene Producers
Committee of the Association of Tar Distillers fixed prices of anthracene until
1965 (R3126, R3127). Prices of creosote oil for hydrogenation were regulated
until 1958 (R3153). Crude tar acids were subject to various pricing agree-
ments between tar distillers and refiners until 1961–1962 (R222, R753, R177,
R178, R179). Finally, sales of briquetting pitch were regulated by a common
selling organization, the Pitch Pool of the Association of Tar Distillers, until
1974 (R3154). All in all, there was a change of competition regime in the long
run in this part of MLH 271.2.
drugs sold in the home market, the industry should probably be seen as
largely competitive.
Paint. Each of the two associations in the paint industry, the Society of
British Paint Manufacturers and the National Federation of Associated Paint,
Colour and Varnish Manufacturers of the U.K., registered a restrictive agree-
ment (R1920, R1896). A third agreement was registered by the British Paint
Advisory Council, which had been set up by the two associations (R385). All
three agreements provided for the maintenance of the manufacturers’ (indi-
vidually set) resale prices, and specified trade discounts for various classes of
buyers and various types of products, quantity discounts, and terms and
conditions of sale. The industry was also listed in CD. The agreements were
canceled in 1959. Although the evidence from PEP suggests that these
agreements may have not been fully effective, the paint industry should be
regarded as one with a change of competition regime.
operation between 1960 and 1968 and was essentially an attempt to resist
an ongoing trend toward higher promotional expenditure in the industry
(R2983). According to MC, Household Detergents (London: H.M.S.O., 1967),
promotional expenditure was still excessive in the mid-1960s, although it
may have declined subsequently. In any case, there is no evidence of any
significant change in the state of competition in the industry between the
1950s and the 1970s.
Finally, according to CD, lead oxide was subject to certain forms of collective
discrimination, such as aggregated rebates, before 1956.
the prices and allocated the supplies of fat and bones for the production of
glue and other products. The products affected by all these arrangements
accounted for about 15% of MLH 279.2 in 1963, but for less than 10% in 1968
and in the 1970s. Because of a substantial change in the definition of the in-
dustry between the 1958 S.I.C. and the 1968 S.I.C., the pre-1963 figures are
not comparable with the post-1963 figures. Hence it is sufficient, for our
present purposes, that the industry can be seen as one without a change in
competition regime after 1963.
Steel products. The register contains more than seventy price-fixing agree-
ments by various associations and groups of steel producers, covering virtu-
ally every product of the steel industry. Some of these products also appear
in CD. Rather than listing the individual agreements, which would be very
356 Appendix A
tedious, I will provide here a summary for the industry as a whole and for its
principal products.3
One of the principal functions of the Iron and Steel Board (the regulatory
body for the steel industry) was to set maximum prices for a wide range of
steel products. For products covered by these controls, the agreements of the
steel producers essentially specified or recommended that the maximum
prices set by the Board be treated as fixed prices. For products free of gov-
ernment controls, the prices were set by the relevant associations. The
agreements typically specified trade discounts and conditions of sale as well.
A few also provided for the reporting of inquiries received from customers,
which could sometimes lead to the sharing of orders.
Most agreements in the steel industry were referred to the Court, and two
were contested by the parties but struck down by the Court. The first
(R1092*) was a pricing agreement by three associations representing manu-
facturers of heavy steel plates, sections, and joists—about 25% of all steel
products (LR 5 RP 33). The second (R1091, R2445*) was between the National
Federation of Scrap Iron, Steel and Metal Merchants and the British Iron and
Steel Federation, and imposed pricing and other restrictions on the market for
steel scrap (LR 4 RP 299). While the vast majority of the registered agreements
were between British firms, the register also contains evidence that the British
Iron and Steel Federation participated in an international restrictive agree-
ment on steel products, at least during 1962–1965 (R2835).
Most agreements were terminated in 1964–1965, but some were abandoned
before 1963. The most important products covered by agreements abandoned
before 1963 were steel castings, steel forgings, and tinplate. All the other
products were covered by agreements canceled after 1963. This distinction is
important in the present context because the steel industry was nationalized
in 1967; hence the only relevant period for our present purposes is the period
1954–1963. Whenever an agreement ended after 1963, the product in ques-
tion can be classified as one that experienced no change in competitive con-
ditions over the relevant period. Whenever an agreement ended before 1963,
however, the product in question must be classified as one with a change in
competition regime.
3. I report here, for the record, the numbers of all the significant restrictive agreements
between steel manufacturers contained in the register. Brief descriptions of those that
were referred to the Court can be found in the Registrar’s reports. Referred to the
Court: R213*, R245*, R256*, R414*, R864*, R913*, R976*, R1079*, R1081*, R1084*,
R1085*, R1092*, R1093*, R1098*, R1100*, R1101*, R1104*, R1105*, R1106*, R1109*,
R1111*, R1112*, R1113*, R1121*, R1122*, R1123*, R1125*, R1128*, R1129*, R1130*,
R1131*, R1243*, R2445*, R2801*, R2802*, R2803*, R4094*. Not referred to the Court:
R861, R862, R863, R1082, R1083, R1086, R1091, R1099, R1102, R1103, R1107, R1108,
R1110, R1114, R1116, R1117, R1118, R1119, R1120, R1124, R1126, R1127, R1202, R1203,
R2430, R2431, R2432, R2835, R3271, R5069, R5074.
Survey of Collusive Agreements in British Industries 357
Pig iron. Several agreements, covering all types of pig iron, were regis-
tered. The Basic Pig Iron Producers Association (R1016, R1097*), the Na-
tional Association of Hematite Pig Iron Makers (R1094*, R1095), the Foundry
Pig Iron Producers Association (R474*) and the Cylinder and Refined Iron
Association (R1087, R1088) all set the prices of their respective products or
recommended that the maximum prices set by the Iron and Steel Board be
treated as fixed prices. Some of these agreements were superseded by others
or ended in the late 1950s. Those that were in force in the early 1960s, cov-
ering more than 90% of total industry sales, were abandoned in 1964–1965.
Thus there was no change in competitive conditions during the period 1954–
1963 in the pig iron industry. After the mid-1960s, the industry was radically
transformed as a result of the nationalization of steel.
358 Appendix A
The register contains restrictive agreements from all three sections of the in-
dustry: smelters, fabricators, and founders. The Federation of Light Metal
Smelters fixed prices for certain types of ingots until 1958 (R192), although
Survey of Collusive Agreements in British Industries 359
Lead manufactures. The Lead Sheet and Pipe Association specified mini-
mum prices for lead sheet and pipes (R1052*). According to CD, it also en-
gaged in certain forms of collective discrimination, including aggregated
rebates, before 1956. The agreement was formally canceled in 1962. Until that
date the association was also a party to an agreement with two smelters re-
garding maximum buying prices of scrap and remelted lead (R803). More-
over, between 1962 and 1965 several members of the association operated an
effective information agreement (R3075). The register also contains a pricing
arrangement of the Association of Lead Sellers, whose membership consisted
of manufacturers and merchants of lead sheet and pipes in Scotland (R908),
and an arrangement among three smelters for the price of scrap and
remelted lead (R1399). Both of these ended in 1962. Lead sheet and pipes
accounted for 15–20% of total sales of lead manufactures in the 1960s.
Other products. The Collapsible Tube Association fixed prices for all types
of collapsible tubes made from nonferrous metals until 1958 (R1258). An
agreement between two leading producers of various ferrous alloys, High
Speed Steel Alloys and Murex, relating to prices, allocation of products, and
some profit-sharing (R1374, R2675), ended in 1962.
Valves. No agreements for valves were registered, but the industry was
mentioned in PEP as allegedly collusive.
The only price agreements for complete machines were the ones registered
by the British Jacquard Engineers’ Association (R515) and the Lace Machine
Builders and Allied Trades Association (R1479). Both were abandoned in
1959–1960. A third agreement, by the British Knitting Machine Builders As-
sociation, comprised only conditions of sale and ended in 1964 (R297). On
the other hand, most parts and accessories of textile machinery were subject
to price fixing until 1959 (R57, R110, R259, R260, R383, R384, R432, R534,
R537, R1427, R1451, R1507). The combined share of jacquard machines and
lace machines in the product category ‘‘complete textile machinery’’ (which
is the one used for the concentration regressions in this book) was probably
around 5%. Even if parts and accessories are taken into account, the collusive
products covered at most 10–15% of total sales revenue of MLH 335. Hence
both the product category ‘‘complete textile machinery’’ and MLH 335 as a
whole can be seen as largely competitive.
on sales promotion, and minimum charges for the renting and maintenance
of typewriters and calculating machines in particular. Most of these restric-
tions were canceled in 1957, but those involving the exchange of information
remained. BT also reports allegations of collective discrimination arrange-
ments in typewriters and typewriter components.
Further details on the industry are provided in CD and PEP (in the latter
case the information comes from a manufacturer of typewriters). It seems
that the original agreement was more restrictive than the one registered, at
least for typewriters: it provided for the collective enforcement of resale prices
and may also have regulated individual prices. On the other hand, PEP also
confirms that many products of the industry were not regulated. The effect of
the 1956 Act on the office machinery industry as a whole was therefore
ambiguous. There is also evidence that there was a change of competition
regime for typewriters in particular.
Ball and roller bearings. No agreement was registered, but both PEP and
BT mention this industry as allegedly collusive.
Pressed Bowl Makers Association. The first three of these ended in 1959, and
R1530* was canceled in 1964.
not regulate the manufacturers’ prices or trade discounts. See the Lloyds’
report and MRPC, Report on the Supply of Dental Goods (London: H.M.S.O.,
1950) for details.
Spectacles and lenses. Prices for spectacles, frames, and lenses were set by
four different associations: the Ophthalmic Prescription Manufacturers As-
sociation (R672), the British Metal Spectacle Manufacturers Association
(R673), the Plastic Spectacle Manufacturers Association (R674), and the Brit-
ish Ophthalmic Lens Manufacturers Association (R675). All four agreements
were abandoned in 1959–1960, leading to a change in competition regime in
this industry.
Switchgear. The register contains two agreements that covered all but the
smaller sizes of switchgear (R689*, R711*). They comprised prices, the report-
ing of inquiries and orders, and design specifications. They were canceled
in 1965–1967. A third agreement, which related to mining-type switchgear,
ended in 1959 (R722). On the other hand, smaller switchgear was subject to
an agreement that involved the reporting of inquiries and conditions of sale,
but did not regulate prices (R695); the arrangement was nevertheless can-
celed in 1960. The register also contains several agreements between manu-
facturers and the Central Electricity Generating Board, regarding prices to be
paid for particular orders of heavy switchgear (R2904, R3077, R3673, R3693).
These continued until the mid-1970s, but it is doubtful whether they restricted
competition to any significant extent. In any case, heavy switchgear ac-
counted for a relatively small part of total industry sales in the 1970s. Thus
the 1956 Act caused a change of competition regime in this industry.
372 Appendix A
Electronic valves, cathode ray tubes, and other active components. The
main association in this industry was the British Radio Valve Manufac-
turers Association, which consisted of producers of electronic valves and
cathode ray tubes for broadcasting or TV receivers. The association members
agreed on common prices for ‘‘maintenance valves’’ (which accounted for
about 10–15% of the total market for valves in the late 1950s) and common
discounts for sales of ‘‘equipment valves’’ to smaller buyers (which accounted
for about 20% of total sales of equipment valves). They also practiced col-
lective exclusive dealing, enforced resale prices, and imposed some restric-
tions on imports and the introduction of new products. See the Lloyds’
report, CD, and especially MRPC, Report on the Supply of Electronic Valves and
Cathode Ray Tubes (London: H.M.S.O., 1957). Following the introduction of
the 1956 Act, the restrictions on common prices and the collective enforcement
of resale price maintenance were withdrawn, and the rest of the agreement
was registered (R313*). It was abandoned in 1959. A pricing agreement by
the Association of Tube Rebuilders, relating to rebuilt cathode ray tubes, also
ended in 1959. The share of products covered by the pricing restrictions
(other than resale price maintenance) in total industry sales revenue was
probably about 20%.
Electric lamps. This industry was tightly regulated in the 1950s. The
agreement of the Electric Lamp Manufacturers Association provided for
common prices, quotas, collective exclusive dealing, aggregated rebates, the
collective enforcement of resale price maintenance, patent pooling, and
restrictions on granting licenses to nonmember firms. There were also inter-
national agreements limiting imports. See the Lloyds’ report, CD, and espe-
cially MRPC, Report on the Supply of Electric Lamps (London: H.M.S.O., 1951).
Some of these restrictions were dropped either as a result of the MRPC in-
quiry or before registration, and so the registered agreement contained only
common trade discounts and the exchange of information on individual
prices (R1548, R2487). These were formally abandoned in 1959. However,
various arrangements in the industry continued throughout the 1960s,
including a price information agreement which led to tacit understandings
about changes in prices and discounts (R3725) and the prenotification of any
new product introductions (R3134, R3283). The industry was officially
investigated for a second time in the late 1960s, and the report concluded
that although competition had increased somewhat since 1956, it was still
not very effective. See MC, Electric Lamps (London: H.M.S.O., 1968). Follow-
ing the investigation, the restrictive arrangements apparently ended. In con-
clusion, the emergence of competition in the electric lamp industry was
gradual and slow, but by the early 1970s the industry had probably experi-
enced a change of competition regime.
Motor bodies, engines, and other parts and accessories of motor vehicles.
The only agreement in the register is a price agreement by the Bus Seat
Frame Association (R957), which was abandoned in 1959. Bus seats are a
very minor product of this industry, which was therefore essentially com-
petitive. Parts and accessories of motor vehicles were also covered by the
collective discrimination arrangements operated by the British Motor Trade
Association; these did not influence the manufacturers’ individual prices or
trade discounts.
The main source of information on this industry is the Report of the Committee
of Inquiry into the Aircraft Industry (London: H.M.S.O., 1965). According to
this report, competition was not very effective, especially for aircraft and
airframes, partly because of the heavy government involvement in the in-
dustry. In particular, competition in aircraft and airframes was being
reduced by government procurement procedures that reflected a wish to
share the available work among the firms. Competition in aero-engines was
more pronounced, at least in the design stage. In the later stages of projects,
competition was affected by the existence of production-sharing arrange-
ments, which the government had approved. These were not registered; in
fact, the only agreement in the register relates to propellers (R41), which is a
minor product of MLH 383. This agreement ended in 1959. On the whole, the
1956 Act had little effect on this industry. Nevertheless, aircraft in particular
was not included in any of the samples used in this book.
(R948), the Circular and Long Saw Association (R977), and the British Hack-
saw Makers Association (R1497, R1670). Some of these agreements were also
mentioned under MLH 390. According to CD and BT, some products of this
industry were subject to collective discrimination arrangements before 1956.
All the price agreements were abandoned in 1959, leading to a change of
competition regime in the hand tool industry.
Wire and wire manufactures of iron and steel. A large number of price
agreements by various associations and groups of wire manufacturers were
registered, covering all types of iron and steel wire and wire manufactures.
Most of these covered specific types of wire or wire manufactures, so it
would be rather tedious to give details on each here.5 According to CD,
several also provided for collective discrimination. Three agreements were
contested in the Court and struck down (see LR 5 RP 146): R867*, by the
Locked Coil Ropemakers Association; R869*, by the Wire Rope Manufac-
turers Association; and R993*, by the Mining Rope Association. Most agree-
ments in this part of MLH 394 were abandoned between 1959 and 1962,
although some, including R867*, R869*, and R993*, were canceled in 1964–
1965. In some cases, information agreements were in operation in the 1960s.
The slow emergence of competition in wire ropes is described in Swann et al.
(1973) and in MC, Wire and Fibre Ropes (London: H.M.S.O., 1973). On the
whole, the 1956 Act caused a change of competition regime in this part of the
wire industry in the long run.
Wire and wire manufactures of copper, brass, and other copper alloys.
The main source of information for this part of MLH 394 is MRPC, Report on
the Supply and Export of Certain Semi-manufactures of Copper and Copper-based
Alloys (London: H.M.S.O., 1955). This report described the collusive activ-
ities of several associations of producers of copper and copper alloy semi-
manufactures, including wire and wire manufactures. See the discussion
under MLH 322. Most of the associations investigated by the MRPC did not
register their agreements. However, the agreements of the Cadmium-Copper
and Bronze Association (R1554) and the High Conductivity Copper Associ-
ation (R508*) were registered and ended in 1959; both covered certain types
of wire. The register also contains two agreements on electric cable wire
(R630 and R634), both abandoned in 1962. To summarize, the 1956 Act
caused a change of competition regime in this part of the wire industry.
5. They are the following: R66, R67, R88, R148, R368, R369, R616, R623, R624, R627,
R628, R629, R631, R632, R633, R635, R636, R637, R638, R659* (by the Mild Steel Wire
Manufacturers Association), R866, R867*, R869*, R945, R993*, R995, R996, R998,
R1004, R1005* (by the Patented Steel Wire Association), R1031* (by a group of manu-
facturers of wire nails), R1067, R1068* (by the Reinforcement Conference), R1072* (by
the British Wire Netting Manufacturers Association), R1190, R1198, R1281, R1303,
R1314, R1377, R1389, R1436, R1845, R1846, R2313, R2981, R3072 (the last of these was
an effective information agreement of the Federation of Reinforcement Fabric Makers).
Some of these may also cover wire and wire products of copper, brass, and other
copper alloys.
384 Appendix A
Until 1960, the Metal Bedstead Association fixed minimum prices for metal
bedsteads, wire and spring mattresses, and other associated products (R1226*).
Moreover, the Bedstead Fittings Association set the prices of bedstead fit-
tings until 1959 (R1251). These products accounted for less than 10% of the
metal furniture industry throughout the 1960s and the 1970s. Office furniture
was subject to some agreements that will be discussed under MLH 472; none
of them was particularly restrictive. Hence the metal furniture industry can
be seen as competitive.
Metal windows and door frames. The British Metal Window Association
fixed the prices of metal windows, door frames, and associated products
(R835, R2618*). In its original form, the agreement also provided for some
allocation of work. See Swann et al. (1973) and MRPC, Report on the Supply of
Standard Metal Windows and Doors (London: H.M.S.O., 1957). The agreement
weakened somewhat in the late 1950s and early 1960s, because the market-
sharing provisions were dropped and the leading producer of metal win-
dows resigned from the association. It was referred to the Court in 1962,
contested by the parties, and allowed to continue (LR 3 RP 198). However, it
was voluntarily abandoned shortly afterward. The register also contains a
price agreement by the Patent Glazing Conference (R836), which ended in
1959. Hence there was a change in competition regime in this industry.
Safes, locks, latches, and keys. The register contains price agreements
by the British Lock and Latch Manufacturers Association (R119) and the
Cylinder Lock Manufacturers Association (R187). They covered keys, locks,
and latches, and were abandoned in 1958–1959. Also relevant is the price
agreement of the National Brassfoundry Association, which covered a wide
range of brass goods, including locks and latches, and was canceled in 1960
(R349). The 1956 Act caused a change of competition regime in this industry.
Coil Spring Association (R994*, R1321) and the Laminated Railway Spring
Association (R1002*, R1003), respectively, were abandoned in 1965. The
combined share of these products in total industry revenue was probably
higher than 50% (at least on the basis the 1968 S.I.C. definition of the industry,
which excludes springs for motor vehicles). Hence there was a change of
competition regime in this industry.
Needles, pins, and other metal small ware. The register contains price
agreements by the Pin and Allied Trades Association (R390), the Metal
Skewer Association (R202), the Machinery Belt Fastener Manufacturers As-
sociation (R2948) and two groups of producers of pins and similar articles
(R2292, R2641). All the agreements were abandoned between 1958 and 1962,
although R390 was replaced for some time by an information agreement. BT
also mentions slide fasteners (zippers) as an allegedly collusive product. The
combined share of the affected products was about 30% of total industry
sales revenue. Several products of the industry were also listed in CD.
taulds had a 50% share in B.N.S. complicates any assessment of the con-
sequences of the Courtaulds-I.C.I. agreement for competition and for the
structure of the industry. Courtaulds was also a party to a number of inter-
national collusive agreements since 1950; these are not, however, relevant for
our present purposes.
In conclusion, only the viscose section of the industry can be regarded as
unambiguously collusive. The share of viscose sales in total industry sales
was less than 50% in the late 1950s, and declined in the 1960s.
Spinning and Doubling on the Cotton and Flax Systems (MLH 412)
Doubled yarn of cotton, glass fibres, and man-made fibres. The Cotton
Yarn Doublers Association set minimum prices for doubled cotton yarn until
1959 (R269*). And the Rayon Staple Spinners and Doublers Association
regulated the price of doubled yarn of man-made staple fibers until 1958
(R1137). However, the price of doubled yarn of man-made continuous fila-
ment apparently was not regulated. The combined share of the products
covered by the two agreements in total industry sales revenue was about
50% in 1958, but only 35% in 1963 and 25% in 1968.
Other products and work done. Until 1959, the Precision Winding Asso-
ciation fixed the charges for electrical-type winding of yarn (R916). A group
Survey of Collusive Agreements in British Industries 389
Woven cotton cloth in the loom state. The register contains three agree-
ments, all of which related to minor products of the industry. An agreement
by the Ventile Fabrics Association of Great Britain, whose membership con-
sisted of spinners, doublers, weavers, and finishers, required members to
deal with each other only (R403); this was discontinued in 1960. A similar
agreement by the Cotton Velvet Council, whose membership consisted of
manufacturers and finishers, was canceled in 1957 (R92). Finally, the Win-
dow Holland Association, which consisted of manufacturers and finishers,
set prices and also operated an exclusive dealing arrangement between the
members until 1959 (R78, R1253). All these were secondary products of the
industry, however, so the influence of these agreements on the industry as
a whole would have been small. According to evidence given by a cotton
weaver and reported in PEP, attempts to reach an agreement between firms
covering the main products of the industry had consistently failed.
Woven cloth of flax or hemp in the loom state. The Association of Flax
Canvas Weavers set the prices of certain categories of flax and/or hemp
canvas (R680). In three separate agreements, certain groups of weavers
agreed upon prices of flax canvas for sale to government departments (R679,
R1487, R1851). These agreements ended in 1959–1960. The products affected
accounted for more than 20% but less than 50% of total industry sales.
operated a market-sharing agreement until 1959 (R2113). After 1959, this was
replaced for some time by a quantity information agreement. Finally, an
agreement by the British Fabric Federation, relating only to conditions of sale
for woven and knitted fabric for wearing apparel, was abandoned in 1967
(R2993*).
Yarn of wool, hair, and man-made fibres spun on the worsted system.
An agreement by the Branded Knitting Wool Association, abandoned in
1960, comprised discounts for hand-knitting wool (R1926). A registered
agreement by the Worsted Spinners Federation, relating only to terms of
contract and settlement discounts, was allowed to continue without a Court
hearing (R1569). According to PEP, however, the federation recommended
minimum prices in the 1950s.
Yarn of wool, hair, and man-made fibres spun on the woollen system.
An agreement by the Scottish Woollen Spinners Association, containing
some minor pricing restrictions, was canceled in 1960 (R218). Neither this
nor the agreement of the Branded Knitting Wool Association described
above (R1926) must have had any significant effect on the industry as a
whole: in both cases the restrictions were minor and the fraction of total in-
dustry sales covered was small. The same can be said about the agreements
Survey of Collusive Agreements in British Industries 391
on yarn for the Harris Tweed industry (see below). Hence the woolen yarn
industry was essentially competitive.
Other products and work done. The British Paper Machine Felt Associa-
tion set the price of felt for industrial machines (R1534). This agreement was
discontinued in 1959, and was replaced by an information agreement. The
Lancashire Mechanical Cloth Manufacturers Association fixed the price of
cloth for industrial machines until 1960 (R917). The Moquette Manufacturers
Association fixed the prices of moquettes for the railways until 1959 (R590).
Finally, the Bradford and District Waste Pullers Association specified charges
for wool shoddy until 1966 (R3032).
The register contains a series of price agreements covering all the main sec-
tions of the jute industry: spinners, weavers, merchants, and importers. In
particular the agreements covered jute yarn (R1340*), jute cloth (R1339*),
hessian (burlap) piece goods (R1337*, R1338*), and jute carpets (R1345*). The
industry was also listed in CD. The first four of these agreements were un-
successfully contested in the Court. All were abandoned in 1963, but they
were replaced by information agreements, which were effective for some
time (see Hart et al. 1973, Swann et al. 1974). There was, however, a change
of competition regime in the long run in the jute industry.
392 Appendix A
Knitted, netted, or crocheted fabrics. The Silk and Rayon Trade Protec-
tion Society, whose membership consisted of weavers, finishers, and mer-
chant converters of various types of fabric, recommended minimum prices
for warp-knitted fabrics and also for certain finishing processes until 1961
(R1637). The affected products accounted for more than 10% but less than
50% of the knitted fabric industry. A price agreement by the Stockinette
Manufacturers Association was abandoned in 1959 (R1448); it covered only a
secondary product of the industry, however. Finally, the agreement of the
British Fabric Federation described under MLH 413 (R2993*) covered prod-
ucts of MLH 417 as well, although it was not particularly restrictive.
Tufted carpets, carpeting, and carpet floor rugs. No evidence of any re-
strictive agreements.
Other carpets, carpeting, and carpet floor rugs, and pile fabric rugs.
Some products that fall into this category were covered by the agreements of
the Federation of British Carpet Manufacturers described above. Moreover,
until 1961 a series of registered agreements by the Association of Manu-
facturers of Mohair and Pile Floor Rugs and Mats provided for specified
trade discounts, the maintenance of manufacturers individual prices, agreed
prices for comparable qualities, and no changes in individual prices without
prior discussion with other members (R557, R1831–R1835).
cotton and electrical tapes until 1962 (R562*). These products accounted
for 10–20% of total sales revenue in this part of MLH 421. In addition, BT
reports allegations of collusion between manufacturers of conveyor belting
and collective exclusive dealing in the brace webbing industry.
Canvas Goods and Sacks and Other Made-Up Textiles (MLH 422.2)
Canvas sacks and bags. Jute sacks and bags were covered by the agree-
ments between producers of hessian (burlap) piece goods described under
MLH 415 (R1337*, R1338*). In addition, the Jute Sack and Bag Manufacturers
Association recommended charges for the sewing, stamping, and bundling
of jute sacks and bags (R1336*). Merchants and importers were also parties
to the pricing arrangements (R1334*, R1335*). All these agreements were
contested in the Court, struck down, and abandoned in 1963. Although they
were replaced for some time by information agreements, the industry expe-
rienced a change of competition regime in the long run as a result of the 1956
Act.
6. Their numbers in the register are as follows: R75, R170, R200, R277, R284, R367,
R393* (by the Dyers and Finishers Association), R402, R682, R830, R838, R942, R979,
R1160, R1245, R1315, R1324, R1531, R1533, R1535, R1567, R1568, R1582, R1606, R1612,
R1637, R1705, R1800, R1801, R2212, R2600.
Survey of Collusive Agreements in British Industries 395
tween 1958 and 1962, although several were initially replaced by information
agreements, some of which lasted until 1970. There was therefore a change of
competition regime in MLH 423 as a result of the 1956 Act.
ations of fellmongers (buyers of sheep hide and skin) (R2690, R2691, R2695).
Under the agreements, each sellers’ association should not sell outside its
territory and each buyers’ association should buy all the sheepskins offered
to it at specified prices. The agreements were abandoned between 1959 and
1961, but they are unlikely to have had any significant impact on competition
between fellmongers. The fellmongery industry was essentially competitive.
in the 1950s by firms in the clothing industry and reported in PEP: according
to one firm, the restrictions on discounts and conditions of sale of the Clothing
Manufacturers Federation were not observed; according to another, the
restrictions on discounts and conditions of sale of the Apparel and Fashion
Industry’s Association were observed, but competition was nevertheless
very intense.
Parts of shoes and work done. The Shoe Tip Association fixed the prices
of heel tips and toe plates until 1961 (R168). An agreement by a group of
manufacturers of cut leather soles provided, until 1959, for specified trade
discounts and aggregated rebates (R1825). Finally, BT mentions allegations
of collective exclusive dealing arrangements in boot and shoe repairing.
(Swann et al. 1974). Bricks were also listed in CD. In summary, there was a
change of competition regime in this industry.
Other products. Most other products of MLH 461 were subject to collusion
in the 1950s. In particular, the register contains price agreements by the Salt
Glazed Conduit Association (R885); the National Clayware Federation, cov-
ering a wide range of clayware goods (R877); the Bridgewater Roofing Tile
Manufacturers Association, covering roofing tiles of clay and fittings (R1646);
the South Eastern Brick and Tile Federation, covering clay bricks and roofing
tiles (R58); the Clay Block Association (R194); the Floor Quarry Association
(R573); and the National Horticultural Pottery Manufacturers Association
(R68). All these agreements ended in 1958–1959. Finally, according to the
case study of the building materials industry in Burn (1958), there was also a
price agreement for earthenware pipes. PEP also reports complaints that the
National Salt Glazed Pipe Manufacturers Association operated a price-fixing
agreement, and the salt glazed pipe industry was also listed in CD and BT.
400 Appendix A
Plate glass. A single producer, Pilkington Brothers, was responsible for the
entire production of plate glass. This firm participated in a number of restric-
Survey of Collusive Agreements in British Industries 401
tive agreements with distributors of plate glass, which were abandoned be-
fore or in 1956. After 1958, it was a party to an international market-sharing
arrangement. See CD and especially MC, A Report on the Supply of Flat Glass
(London: H.M.S.O., 1968). Unprocessed plate glass is the main product of
this part of MLH 463. Another important product is processed plate glass.
The register contains a price agreement by the Plate Glass Association, whose
membership consisted of processors and dealers in plate glass (R821*). This
agreement was abandoned in 1959, and was replaced by an information
agreement, which ended in 1965. The register also contains price-fixing agree-
ments by the Glass Benders Association (R837*), the Scottish Glass Mer-
chants and Glaziers Association (R1188) and the N. Ireland Plate Glass
Association (R827), all of which were cancelled in 1958–1960. R837* and
R1188 were replaced by information agreements, which lasted until the mid-
1960s. The share of processed plate glass in the total sales revenue of the
plate glass industry was 15–30% in the 1960s.
Glass containers. The agreement of the Glass Bottle Association set mini-
mum prices for glass containers (R660*, R1681, R1682). It was contested in the
Court and struck down in 1961 (LR 2 RP 345). Swann et al. (1973) provide
details on this case. The 1956 Act caused a change of competition regime in
this industry.
The agreement of the Cement Makers Federation (R77*), which provided for
common prices and aggregated rebates, was contested in the Court in 1961,
and the price restrictions were allowed to continue (see LR 2 RP 241 and the
report of the Director General of Fair Trading for 1974). The agreement was
still in operation in the late 1970s. Swann et al. (1973) describe the evolution
of the cement industry from the 1950s to the early 1970s. Collusion continued
in the industry throughout the period examined in this book.
Lime and whiting. The register contains price agreements by the Southern
Lime Association (R1248), the South West of England Lime Association
(R2268), and the British Whiting Federation (R1259). The lime industry was
also listed in CD. All agreements were abandoned in 1959, leading to a
change of competition regime in this industry.
least to some extent, during the 1960s: a regional agreement that was in op-
eration until 1967 (R2818), and an agreement by some of the largest firms in
the industry with respect to a particular tender for a government contract in
1966–1967 (R3106). This evidence suggests that it may be best to classify this
industry as ambiguous.
Other products and work done. The Kiln Owners Association set a sched-
ule of charges for kiln drying of timber and plywood until 1958 (R520).
The National Bedding Federation, which was a member of the British Fur-
niture Trade Confederation, had an agreement similar to the agreement of
the latter (see above, under MLH 472). In particular, members were to deal
only with one another and distributors were to maintain the manufacturers’
retail prices (R1686*). A separate agreement of the federation provided for
quantity discounts to be allowed by manufacturers on mattresses sold directly
to hotels (R173*). Both agreements ended in 1962–1963. A price agreement
by the National Association of Upholstery Fibre Processors, covering uphol-
stery and mattress pads, was abandoned in 1958 (R671). This was a second-
ary product of the industry, however. Thus the bedding and soft furnishings
industry was largely competitive.
The Shive Manufacturers Association fixed prices until 1959 (R2304). This is
a minor product of the industry. Neither this agreement, nor some local price
agreements on packing cases, abandoned at various dates between 1957 and
1967, could have had any significant effect on competition for the industry as
a whole.
7. They are the following: R497, R498, R499* (by the Association of Makers of Wood
Free Papers), R500, R501, R502, R503, R504, R505, R506, R507, R556* (by the Associa-
tion of Corrugated Paper Makers), R923, R978, R1237, R1238, R1239, R1240, R1241,
R1265, R1266, R1267, R1268, R1282, R1328, R1361, R1364, R1365, R1366, R1367, R1368,
R1370, R1456, R1501, R1735, R2224, R2635.
408 Appendix A
A price agreement by the British Paper Bag Federation fixed prices for paper
bags until 1965 (R1235). Moreover, the Royal Hand and Grocery Bags Asso-
ciation set prices for soda bags and sugar bags until 1960 (R1369). Paper bags
accounted for about 30% of total sales revenue in the whole of MLH 482.2
and for about 45% of total sales revenue of the product category ‘‘paper bags
and sacks’’ in the 1960s.
Survey of Collusive Agreements in British Industries 409
Other products. The Association of Lace Paper Makers (R495), the Associ-
ation of Makers of Paper Serviettes (R496), the Society of Crepe Paper Makers
(R491), and the Federation of Paper Tube Manufacturers (R183) fixed prices
for their respective products until 1959. A price agreement by the Vulcanised
Fibre and Leatheroid Association ended in 1958 (R1076).
R2721, R2722, R2805*, R2909, R2965*). Most of these agreements were dis-
continued in 1964–1966, although in a few cases agreements on price
increases were made as late as 1970 (R3590*). Several other regional agree-
ments provided for the prenotification of price increases and restrictions on
sales promotion; in fact, in most cases the firms involved were required not
to change the prices of their newspapers without prior discussion with the
other parties (R2611*, R2394–R2420). All these ended in 1960–1962. Other
agreements between publishers, mostly regional, comprised conditions of
sale (including a requirement not to sell newspapers before a certain time of
the day), sales promotion, or restrictions on the entry of retailers, and ended
at various dates between 1957 and 1964 (R1653*, R2003, R2004, R2299,
R2300, R2387, R2388, R2422, R2610, R2612). Finally, a series of agreements
between publishers regulating the distribution of newspapers and providing
for agreed trade discounts and conditions of sale were abandoned in the late
1960s or the early 1970s (R2385*, R2386, R2604, and many others similar in
form to those three).
In summary, the industry was probably collusive in the 1950s. It experi-
enced a change of competition regime as a result of the 1956 Act, although
the effect of the Act was probably not fully realized until the 1970s.
R2608*). PEP also mentions the Greeting Card and Calendar Association as
potentially collusive.
competition regime in the tire industry in the long run. According to Swann
et al. (1973), competition between manufacturers was intense in the early
1970s.
Other rubber goods and work done. Price agreements by the Rubber
Proofers Association (R914*), the Screw Stopper Makers Association (R1355),
and a group of producers of insulated, varnished, nonadhesive cambric or
silk cloth and tape (R710) were abandoned in 1957–1959.
Lead Pencils. The Pencil Makers Conference fixed the prices of cheaper
pencils, although not the prices of the more expensive varieties (R1426). This
may be seen as a minimum-price agreement. Hence the pencil industry
probably experienced a change of competition regime as a result of the 1956
Act.
Other products. The Briar Pipe Trade Association set minimum prices until
1959 (R600). An agreement by a group of producers of peat moss litter, relat-
ing to prices and the allocation of work, ended in 1961 (R799). Finally, BT
reports allegations about a restrictive association in casein buttons, although
it does not specify the types of restrictions. This may refer to an exclusive
dealing arrangement between sellers of casein plastic and producers of
casein buttons, also reported in CD.
Appendix B: Datasets
Table B1
416
Data set for chapter 4
Industry Year C5 SS DS K/N K/L CHANGE
Clay, brick earth, marl, and shale 1958 0.726 13066 12745 0.51 4.91 1
1963 0.821 17827 15996 0.74 7.21 1
1968 0.910 32001 32001 1.25 11.72 1
1975 0.966 33165 26369 1.91 17.21 1
Iron ore and ironstone 1958 0.789 13502 10856 1.49 7.68 0
1963 0.884 12748 11475 1.88 11.26 0
White flour for breadmaking 1958 0.715 121751 123922 0.77 4.60 1
1963 0.792 113211 119913 1.14 6.45 1
1968 0.808 110439 110439 1.56 9.50 1
1975 0.895 135726 108612 2.29 13.20 1
Bread 1958 0.315 184559 238086 0.46 2.57 1
1963 0.714 214081 245906 0.58 2.47 1
1968 0.773 236006 236006 0.86 2.93 1
1975 0.823 207525 212485 0.92 3.42 1
Bacon and ham, cured and smoked 1958 0.308 90730 86486 0.21 1.43 1
1963 0.474 74194 78009 0.31 1.70 1
1968 0.569 67902 67902 0.55 2.44 1
1975 0.532 102384 74775 0.66 3.60 1
Appendix B
Sausages and sausage meat 1958 0.505 44575 44238 0.21 1.43 0
1963 0.522 52110 52719 0.31 1.70 0
1968 0.562 59218 59218 0.55 2.44 0
1975 0.565 62598 58139 0.66 3.60 0
Preserved fruit, other than marmalade and 1958 0.308 23915 19305 0.46 1.95 0
Data Sets
jams 1963 0.377 20428 19915 0.86 2.99 0
1968 0.457 23267 23267 1.28 4.28 0
1975 0.444 18855 18775 2.12 6.33 0
Vegetable and seed oils 1963 0.823 85828 84149 1.11 7.75 0
1968 0.843 80519 80519 1.36 10.85 0
1975 0.869 122108 86560 2.28 16.85 0
Fish and marine animal oils 1958 0.949 12364 5893 0.93 6.31 0
1963 0.928 14080 9922 1.11 7.75 0
1968 0.929 14129 14129 1.36 10.85 0
1975 0.913 20553 12343 2.28 16.85 0
Oilseed cake and meal, meat meal, 1963 0.804 27226 26755 1.11 7.75 0
bonemeal, etc. 1968 0.726 21508 21508 1.36 10.85 0
1975 0.817 34060 34840 2.28 16.85 0
Malt 1958 0.534 24538 17595 0.81 3.78 0
1963 0.551 30607 25993 1.36 5.31 0
1968 0.651 30347 30347 3.03 8.41 0
1975 0.724 62701 54227 7.20 16.76 0
Coke 1958 0.619 162360 154766 6.02 21.34 0
1963 0.653 153464 149435 9.73 28.18 0
Steel ingots 1958 0.842 55246 45286 7.54 11.35 0
1963 0.766 48954 45106 12.15 17.26 0
Steel blooms, billets, and slabs 1958 0.722 258111 223065 7.54 11.35 0
1963 0.696 203506 187510 12.15 17.26 0
417
Table B1 (continued)
418
Industry Year C5 SS DS K/N K/L CHANGE
Steel arches, etc. 1958 0.636 27664 23575 7.54 11.35 0
1963 0.692 21136 19474 12.15 17.26 0
Hot rolled coil, etc. 1958 0.688 135593 116048 7.54 11.35 0
1963 0.728 174121 160434 12.15 17.26 0
Steel plates, uncoated, 3 mm thick and 1958 0.789 159323 131141 7.54 11.35 0
over 1963 0.724 130471 120215 12.15 17.26 0
Steel plates and sheets, uncoated, under 3 1958 0.911 173290 145342 7.54 11.35 0
mm thick 1963 0.942 166268 153199 12.15 17.26 0
Wrought tubes, manipulated, fabricated, 1958 0.858 33584 32245 2.25 4.92 1
etc. 1963 0.733 30886 30007 3.21 7.29 1
1968 0.793 35437 35437 3.68 8.03 1
Pig iron 1958 0.547 125126 107090 0.61 3.57 0
1963 0.673 97637 89962 0.97 4.82 0
Iron castings 1958 0.250 225786 222694 0.61 3.57 1
1963 0.329 266582 259289 0.97 4.82 1
1968 0.422 272333 272333 1.21 5.65 1
Aluminium and aluminium alloys, and 1958 0.504 177881 162181 1.39 4.22 1
manufactures thereof 1963 0.542 218852 220888 1.55 5.02 1
1968 0.509 282354 282354 2.22 7.32 1
Appendix B
1975 0.541 299943 316371 4.24 15.30 1
Copper and manufactures thereof 1958 0.746 149102 265497 1.23 5.05 1
Data Sets
1963 0.741 179616 307212 1.35 5.80 1
1968 0.798 254943 254943 1.64 6.71 1
1975 0.830 150489 239120 1.94 7.76 1
Nickel and nickel alloys, and 1963 0.954 55743 66342 1.83 7.17 0
manufactures thereof 1968 0.970 63786 63786 2.13 9.81 0
Tin and tin alloys, and manufactures 1958 0.913 42598 60224 1.74 6.99 0
thereof 1963 0.893 36996 45224 1.83 7.17 0
1968 0.916 60136 60136 2.13 9.81 0
1975 0.978 41179 38656 2.69 15.48 0
Zinc and zinc alloys, and manufactures 1958 0.686 28233 35841 1.74 6.99 0
thereof 1963 0.690 37667 41810 1.83 7.17 0
1968 0.713 44085 44085 2.13 9.81 0
1975 0.734 29847 18562 2.69 15.48 0
Track-laying tractors 1958 0.902 8315 9536 1.29 3.43 0
1963 0.980 18749 18488 1.58 3.64 0
Lifts, escalators, and passenger conveyors 1958 0.625 12422 12452 0.29 1.41 1
1963 0.616 15513 14986 0.33 1.56 1
1968 0.641 20256 20256 0.41 1.91 1
1975 0.751 29538 28427 0.53 2.44 1
Mining machinery 1958 0.362 36573 36848 0.51 2.09 0
1963 0.422 43586 42979 0.56 2.38 0
1968 0.591 38306 38306 0.70 2.81 0
Space-heating, ventilating, and air- 1958 0.233 45635 45978 0.37 2.13 0
419
conditioning equipment (1958 S.I.C.) 1963 0.241 78532 77438 0.36 2.06 0
Table B1 (continued)
420
Industry Year C5 SS DS K/N K/L CHANGE
Fans and ventilating units, industrial (1968 1963 0.737 15938 15716 0.36 2.06 0
S.I.C.) 1968 0.604 19641 19641 0.41 2.41 0
1975 0.566 21960 20764 0.50 3.31 0
Bottling, packing, canning, packeting, and 1963 0.553 17937 17687 0.49 2.23 0
labeling machinery 1968 0.567 21958 21958 0.59 2.82 0
1975 0.592 22718 22118 0.71 3.84 0
Automatic slot machines 1963 0.903 2604 2567 0.49 2.23 0
1968 0.721 15305 15305 0.59 2.82 0
1975 0.724 8305 8173 0.71 3.84 0
Pulpmaking, papermaking, and 1958 0.840 10981 11063 0.42 2.07 0
boardmaking machinery 1963 0.938 22077 21770 0.49 2.23 0
1968 0.833 16208 16208 0.59 2.82 0
1975 0.825 13125 12916 0.71 3.84 0
Ordnance and small arms 1958 0.712 51490 51877 1.61 2.58 0
1963 0.625 40566 40001 2.42 3.45 0
1968 0.766 55543 55543 3.16 5.08 0
Bearings, other than ball and roller 1958 0.847 16338 16461 0.76 3.48 0
bearings, and bushes 1963 0.739 26617 26246 0.73 3.62 0
1968 0.866 26012 26012 0.92 3.95 0
Appendix B
1975 0.822 30512 28849 1.07 5.36 0
Gaskets and jointings 1963 0.762 8337 8221 0.73 3.62 0
1968 0.806 16709 16709 0.92 3.95 0
1975 0.904 11250 10636 1.07 5.36 0
Electric cables, other than for 1958 0.533 106228 114259 3.00 4.28 1
Data Sets
telecommunication 1963 0.680 137090 154163 2.84 4.68 1
1968 0.818 149872 149872 3.75 5.42 1
Electric light fittings, wiring and 1958 0.381 41734 37492 0.49 1.72 1
accessories, etc. 1963 0.380 61029 60668 0.63 2.03 1
1968 0.473 67898 67898 0.98 2.68 1
1975 0.470 84145 81542 1.43 4.13 1
Trailers and caravans 1958 0.370 24164 24469 3.47 6.46 1
1963 0.435 36093 36472 3.49 5.81 1
1968 0.442 71277 71277 3.71 6.49 1
1975 0.346 79799 73079 3.78 6.78 1
Hand tools and implements 1958 0.286 30391 32426 0.29 1.92 1
1963 0.333 38833 39194 0.35 2.34 1
1968 0.376 39526 39526 0.41 2.73 1
1975 0.522 45951 39840 0.55 3.62 1
Iron and steel wire 1958 0.561 73452 70269 0.60 3.42 1
1963 0.577 80919 78527 0.71 4.15 1
1968 0.573 82320 82320 0.95 5.27 1
1975 0.655 87046 68175 1.20 6.26 1
Manufactures of iron and steel wire 1958 0.417 59268 59338 0.60 3.42 1
1963 0.445 76581 77202 0.71 4.15 1
1968 0.512 86233 86233 0.95 5.27 1
1975 0.483 131816 122915 1.20 6.26 1
Wire of brass and other copper alloys, and 1958 0.690 11496 17529 0.60 3.42 1
manufactures thereof 1963 0.752 14570 18442 0.71 4.15 1
421
1968 0.829 18604 18604 0.95 5.27 1
Table B1 (continued)
422
Industry Year C5 SS DS K/N K/L CHANGE
Copper wire and manufactures thereof 1958 0.698 41459 72366 0.60 3.42 1
1963 0.724 23503 43499 0.71 4.15 1
1968 0.869 38609 38609 0.95 5.27 1
Cans, metal boxes, and other small metal 1958 0.869 91347 80517 0.84 2.98 0
containers 1963 0.904 109910 104299 1.47 4.07 0
1968 0.908 126711 126711 2.24 6.04 0
Precious metals, refined 1963 0.954 83585 105953 0.36 2.56 0
1968 0.965 138771 138771 0.68 4.10 0
Jewellery and plate (1958 S.I.C.) 1958 0.218 25559 22731 0.30 2.23 0
1963 0.263 34946 29945 0.36 2.56 0
Jewellery and plate (1968 S.I.C.) 1963 0.336 23130 19820 0.36 2.56 0
1968 0.495 27497 27497 0.68 4.10 0
1975 0.441 63290 49942 0.82 4.95 0
Metal furniture 1958 0.241 42492 41095 0.29 2.47 0
1963 0.276 57570 57230 0.27 1.87 0
1968 0.225 66030 66030 0.33 2.24 0
1975 0.249 64974 63022 0.42 3.23 0
Drop forgings of steel and iron, and steel 1963 0.583 103497 89132 0.41 2.40 1
stampings and pressings 1968 0.612 104139 104139 0.50 2.99 1
Appendix B
1975 0.650 115994 95893 0.58 3.72 1
Metal hollowware, domestic and 1958 0.252 61627 68700 0.33 2.14 1
industrial 1963 0.335 66378 71636 0.41 2.40 1
1968 0.327 67972 67972 0.50 2.99 1
1975 0.412 60195 54461 0.58 3.72 1
Metal windows, metal door frames, etc. 1958 0.705 26531 27460 0.33 2.14 1
Data Sets
1963 0.583 30510 30294 0.41 2.40 1
1968 0.610 33217 33217 0.50 2.99 1
1975 0.509 52420 34309 0.58 3.72 1
Safes, locks, latches, keys, etc. 1958 0.687 9517 12377 0.33 2.14 1
1963 0.764 15224 18296 0.41 2.40 1
1968 0.743 18948 18948 0.50 2.99 1
1975 0.753 21861 19916 0.58 3.72 1
Single yarn of cotton, glass fibres, and 1958 0.319 188803 159023 0.74 3.39 1
man-made fibres 1963 0.372 134557 132227 1.22 5.22 1
1968 0.503 123935 123935 1.81 7.46 1
1975 0.756 82616 78395 2.58 11.05 1
Finished thread for sewing, embroidery, 1958 0.851 26988 22009 0.74 3.39 0
etc. 1963 0.818 23737 22278 1.22 5.22 0
1968 0.879 27491 27491 1.81 7.46 0
1975 0.879 27080 25558 2.58 11.05 0
Woven cotton cloth in the loom state 1958 0.116 157737 145988 0.46 3.25 0
1963 0.193 83641 86041 0.75 4.80 0
1968 0.313 52050 52050 1.08 7.22 0
1975 0.437 27205 25111 1.44 9.85 0
Woven cloth of man-made fibres in the 1958 0.211 99105 78882 0.46 3.25 1
loom state 1963 0.358 91010 81405 0.75 4.80 1
1968 0.519 84540 84540 1.08 7.22 1
1975 0.636 59521 60036 1.44 9.85 1
423
Table B1 (continued)
424
Industry Year C5 SS DS K/N K/L CHANGE
Wool tops or slubbings 1958 0.301 121031 96334 0.39 2.49 0
1963 0.340 120401 86844 0.50 2.98 0
1968 0.547 86161 86161 0.73 4.37 0
1975 0.642 44690 52171 1.06 6.97 0
Yarn of wool, hair, and man-made fibres 1958 0.237 52506 46333 0.39 2.49 0
spun on the woollen system 1963 0.230 72396 65786 0.50 2.98 0
1968 0.339 68272 68272 0.73 4.37 0
1975 0.370 59914 63038 1.06 6.97 0
Woven worsted fabrics 1958 0.173 109476 100736 0.39 2.49 0
1963 0.267 100029 93473 0.50 2.98 0
1968 0.311 79842 79842 0.73 4.37 0
1975 0.438 50733 50160 1.06 6.97 0
Woven woollen fabrics 1958 0.177 103410 99048 0.39 2.49 0
1963 0.151 88506 86576 0.50 2.98 0
1968 0.240 79218 79218 0.73 4.37 0
1975 0.268 44284 43274 1.06 6.97 0
Jute yarn, cloth, and other manufactures of 1958 0.508 36494 42969 0.94 4.09 1
jute 1963 0.592 40097 43219 1.07 4.01 1
1968 0.689 38728 38728 1.24 4.97 1
1975 0.734 12990 15466 2.27 9.30 1
Appendix B
Rope, twine, net, and manufactures 1958 0.477 29692 26886 0.34 2.30 1
thereof 1963 0.648 33161 27166 0.45 2.86 1
1968 0.732 24197 24197 0.77 4.24 1
1975 0.821 21168 17356 1.30 6.82 1
Knitted, netted, or crocheted goods: 1958 0.256 40530 31566 0.60 4.26 0
Data Sets
underwear, shirts, and nightwear 1963 0.395 46060 37270 0.68 4.23 0
1968 0.531 48760 48760 0.86 4.92 0
1975 0.504 41022 46169 0.99 5.81 0
Knitted, netted, or crocheted goods: other 1958 0.150 88309 93221 0.60 4.26 0
clothing 1963 0.212 113456 105789 0.68 4.23 0
1968 0.332 139594 139594 0.86 4.92 0
1975 0.402 125522 163748 0.99 5.81 0
Woven carpets, carpeting, and carpet floor 1958 0.412 80749 77717 0.83 2.96 1
rugs 1963 0.423 84798 80025 0.91 3.16 1
1968 0.575 81014 81014 1.24 3.80 1
Bedding other than wool blankets 1963 0.446 28247 27350 0.08 0.76 0
1968 0.513 37014 37014 0.14 1.30 0
1975 0.555 49565 49447 0.24 2.37 0
Other made-up household textiles 1963 0.271 28631 27722 0.08 0.76 0
1968 0.266 26181 26181 0.14 1.30 0
1975 0.474 25377 21230 0.24 2.37 0
Canvas goods and sacks 1958 0.252 33245 26594 0.10 0.94 1
1963 0.280 27742 24382 0.12 1.15 1
1968 0.249 21177 21177 0.14 1.43 1
1975 0.313 17515 16040 0.14 1.73 1
Asbestos manufactures other than asbestos 1958 0.774 41279 37379 1.81 3.28 1
cement goods 1963 0.824 49883 47653 2.12 3.59 1
1968 0.862 58645 58645 2.89 4.45 1
425
1975 0.867 52739 60991 4.33 6.23 1
Table B1 (continued)
426
Industry Year C5 SS DS K/N K/L CHANGE
Leather, undressed 1958 0.382 24405 25969 0.20 1.98 0
1963 0.453 16743 17306 0.24 2.34 0
1968 0.556 13845 13845 0.36 3.28 0
1975 0.763 11320 12061 0.49 4.50 0
Leather, dressed 1963 0.225 63802 66350 0.24 2.34 0
1968 0.217 75429 75429 0.36 3.28 0
1975 0.339 62049 70324 0.49 4.50 0
Fellmongery 1958 0.461 8928 7496 0.20 1.98 0
1963 0.427 16260 11909 0.24 2.34 0
1968 0.422 10523 10523 0.36 3.28 0
1975 0.506 8753 9210 0.49 4.50 0
Manufactures of leather or leather 1958 0.213 19570 20523 0.04 0.42 0
substitutes 1963 0.208 26438 26747 0.06 0.58 0
1968 0.203 28918 28918 0.07 0.70 0
1975 0.245 31140 32183 0.09 1.01 0
Skins and furs, dressed, etc., and 1963 0.377 11243 10495 0.10 0.82 0
manufactures thereof 1968 0.495 10252 10252 0.14 1.13 0
1975 0.635 7938 9554 0.16 1.59 0
Weatherproof outerwear 1963 0.223 50884 45994 0.08 0.74 0
Appendix B
1968 0.306 42361 42361 0.10 0.93 0
1975 0.298 40485 50254 0.12 1.08 0
Men’s and boys’ tailored outerwear 1958 0.350 148333 129315 0.07 0.34 0
Data Sets
1963 0.380 158776 146296 0.09 0.47 0
1968 0.377 158184 158184 0.13 0.67 0
1975 0.283 144680 168807 0.17 0.99 0
Women’s and girls’ tailored outerwear 1958 0.152 82374 75103 0.05 0.53 0
1963 0.167 86343 82575 0.06 0.59 0
1968 0.201 85066 85066 0.09 0.78 0
1975 0.177 99846 123790 0.10 0.87 0
Industrial and heavy overalls, aprons, and 1958 0.241 18897 18872 0.06 0.47 0
jeans 1963 0.261 22866 22522 0.08 0.65 0
1968 0.274 22717 22717 0.12 0.91 0
1975 0.326 33242 33321 0.14 1.11 0
Women’s and girls’ light outerwear 1958 0.082 83327 75540 0.03 0.40 0
1963 0.129 76555 72727 0.05 0.52 0
1968 0.166 95702 95702 0.08 0.74 0
1975 0.160 133928 187785 0.09 0.92 0
Lingerie other than corsetry and brassieres 1958 0.202 35242 33001 0.03 0.40 0
1963 0.265 34980 33304 0.05 0.52 0
1968 0.414 37978 37978 0.08 0.74 0
1975 0.453 30193 38132 0.09 0.92 0
Infants’ wear 1958 0.177 16519 15026 0.03 0.40 0
1963 0.153 17190 16046 0.05 0.52 0
1968 0.297 20488 20488 0.08 0.74 0
1975 0.288 17143 25685 0.09 0.92 0
427
Table B1 (continued)
428
Industry Year C5 SS DS K/N K/L CHANGE
Gloves, other than sport, knitted, and 1958 0.284 11629 11687 0.06 0.67 0
rubber gloves 1963 0.327 12894 12642 0.06 0.73 0
1968 0.414 10734 10734 0.08 0.92 0
1975 0.485 9125 9222 0.13 1.47 0
Tiles, other than of precast concrete and 1958 0.648 12642 12407 0.40 1.59 0
brick earth 1963 0.702 16120 15331 0.54 1.98 0
1968 0.936 17562 17562 0.73 2.48 0
1975 0.932 17522 18289 0.94 2.97 0
Sanitary ware of earthenware 1958 0.641 6972 6327 0.40 1.59 1
1963 0.663 11399 11138 0.54 1.98 1
1968 0.951 12733 12733 0.73 2.48 1
1975 0.939 14665 15067 0.94 2.97 1
China, earthenware, etc., domestic and 1958 0.274 37170 43571 0.40 1.59 1
ornamental 1963 0.317 40657 43386 0.54 1.98 1
1968 0.484 46653 46653 0.73 2.48 1
1975 0.576 72952 69921 0.94 2.97 1
Electrical ware of porcelain, earthenware, 1958 0.769 8289 8289 0.40 1.59 1
or stoneware 1963 0.808 11521 11092 0.54 1.98 1
1968 0.857 12014 12014 0.73 2.48 1
1975 0.754 11610 10957 0.94 2.97 1
Appendix B
Glass containers 1958 0.632 51899 47427 0.83 2.65 1
1963 0.696 60251 60456 1.41 3.68 1
1968 0.873 70260 70260 2.13 5.45 1
1975 0.876 84865 77991 2.89 8.70 1
Cement 1958 0.855 77393 71133 2.71 10.58 0
Data Sets
1963 0.890 90038 84136 3.79 14.40 0
1968 0.890 101579 101579 6.26 23.88 0
1975 0.930 126930 111057 6.83 25.60 0
Asbestos cement goods 1958 0.983 19849 19923 0.46 3.39 1
1963 0.942 24842 25524 0.63 4.58 1
1968 0.986 24510 24510 0.81 6.12 1
1975 0.982 17075 15740 1.19 8.86 1
Plywood and veneers 1963 0.336 17458 20244 0.16 1.32 0
1968 0.444 35998 35998 0.18 1.52 0
Textured boards and wood chipboard 1963 0.555 8768 8538 0.16 1.32 0
1968 0.544 19395 19395 0.18 1.52 0
Builders’ woodwork and prefabricated 1958 0.192 51985 58608 0.13 1.17 1
building structures of timber 1963 0.184 83526 89132 0.16 1.32 1
1968 0.240 125719 125719 0.18 1.52 1
1975 0.329 126038 125619 0.27 2.66 1
Office, school, and other furniture 1958 0.225 31472 31313 0.10 0.78 0
1963 0.243 37553 37858 0.12 0.93 0
1968 0.257 40945 40945 0.19 1.38 0
1975 0.284 43702 38301 0.33 2.36 0
Wooden containers and baskets 1958 0.116 33083 32711 0.09 0.91 0
1963 0.162 36207 36717 0.10 1.05 0
1968 0.176 36502 36502 0.13 1.33 0
1975 0.217 33061 30625 0.15 1.80 0
429
Table B1 (continued)
430
Industry Year C5 SS DS K/N K/L CHANGE
Paper, uncoated 1958 0.526 250871 218468 2.74 8.72 1
1963 0.566 261650 260101 3.23 9.61 1
1968 0.548 262225 262225 4.91 13.74 1
Paper and board, coated 1958 0.407 39023 31856 2.74 8.72 1
1963 0.536 52930 46615 3.23 9.61 1
1968 0.632 62511 62511 4.91 13.74 1
Paper and board, oiled, waxed, and other 1958 0.519 18681 16651 2.74 8.72 1
waterproof wrappings 1963 0.603 16882 16401 3.23 9.61 1
1968 0.647 17131 17131 4.91 13.74 1
Boxes and cartons of paper and cardboard 1958 0.314 78812 81573 0.49 3.21 0
1963 0.303 98159 95149 0.62 3.64 0
1968 0.387 113954 113954 0.81 4.27 0
1975 0.385 142961 116943 1.10 6.30 0
Fibreboard packing cases 1958 0.736 61670 52442 0.49 3.21 0
1963 0.655 91999 84586 0.62 3.64 0
1968 0.675 119657 119657 0.81 4.27 0
1975 0.631 149089 114461 1.10 6.30 0
Rubber cellular products 1958 0.863 12392 10971 1.27 3.80 0
1963 0.774 11688 10815 1.54 4.06 0
Appendix B
Carbons and other office machinery 1958 0.633 10662 9709 0.29 1.63 0
requisites 1963 0.642 14472 13871 0.43 2.33 0
1968 0.749 17506 17506 0.56 3.41 0
1975 0.792 17211 18894 0.81 5.56 0
Components, accessories, and 1963 0.120 48989 37990 0.33 2.24 0
Data Sets
semimanufactured goods of plastics 1968 0.164 84334 84334 0.55 3.60 0
1975 0.136 113173 125688 0.82 5.96 0
Domestic, catering, and furnishing goods 1963 0.235 19127 14833 0.33 2.24 0
of plastics 1968 0.299 31250 31250 0.55 3.60 0
1975 0.254 45655 52945 0.82 5.96 0
Packaging materials of plastics 1963 0.443 37290 28918 0.33 2.24 0
1968 0.411 75053 75053 0.55 3.60 0
1975 0.302 116905 139264 0.82 5.96 0
Musical instruments 1958 0.454 5356 5929 0.16 1.38 0
1963 0.501 6643 6642 0.21 1.63 0
Notes: C5 is the five-firm sales concentration ratio.
SS is sales revenue (in £1000) deflated by the general producer price index.
DS is sales revenue (in £1000) deflated by industry-specific producer price indices.
K is the value of capital stock of the corresponding MLH industry in 1980 prices (in £ million).
L is employment of the corresponding MLH industry (in 1000).
N is the number of plants with at least 25 employees of the corresponding MLH industry.
CHANGE takes the value 1 for an industry with a change in competition regime and 0 otherwise.
431
Table B2
432
Data set for the concentration regressions of chapter 5
Appendix B
Milk powder 1958 0.737 14686 12692 0.62 3.65 0.25 1 1
1963 0.889 15051 13633 0.74 4.74 0.55 1 1
1968 0.847 18692 18692 0.92 5.81 0.85 1 1
1975 0.753 38195 23339 1.35 8.27 0.85 1 1
Ice cream and ice lollies 1958 0.942 22392 18076 0.62 3.65 0.55 2 0
Data Sets
1963 0.931 14079 15895 0.74 4.74 0.85 2 0
1968 0.912 19276 19276 0.92 5.81 0.85 2 0
1975 0.909 49483 30115 1.35 8.27 0.55 2 0
Cocoa products 1958 0.732 176041 163309 0.84 2.42 0.55 2 1
1963 0.823 171639 166754 1.40 3.40 0.85 2 1
1968 0.834 180991 180991 2.22 4.64 0.85 2 1
1975 0.836 218224 184031 3.27 6.72 0.85 2 1
Marmalade and jams 1958 0.629 36118 34021 0.46 1.95 0.55 2 0
1963 0.729 34564 33391 0.86 2.99 0.55 2 0
1968 0.756 31488 31488 1.28 4.28 0.55 2 0
1975 0.689 30137 25043 2.12 6.33 0.85 2 0
Vegetables, etc., preserved 1958 0.649 45235 35897 0.46 1.95 0.55 1 0
(1958 S.I.C.) 1963 0.672 47825 42912 0.86 2.99 0.55 1 0
Vegetables, etc., preserved 1963 0.653 65177 58482 0.86 2.99 0.55 1 0
(1968 S.I.C.) 1968 0.667 73798 73798 1.28 4.28 0.85 1 0
1975 0.703 101902 95830 2.12 6.33 0.85 1 0
Vegetables, quick-frozen 1958 0.986 7340 8287 0.46 1.95 0.85 1 0
1963 0.933 24913 23826 0.86 2.99 0.85 1 0
1968 0.971 35911 35911 1.28 4.28 0.85 1 0
1975 0.817 58919 62240 2.12 6.33 0.85 1 0
Pickles, sauces, and relishes 1963 0.680 28206 27342 0.86 2.99 0.25 2 0
1968 0.715 30293 30293 1.28 4.28 0.55 2 0
1975 0.679 41689 42479 2.12 6.33 0.55 2 0
433
Table B2 (continued)
434
Industry Year C5 SS DS K=N K=L TVCLASS ADTYPE CHANGE
Soups 1958 0.911 25732 21650 0.46 1.95 0.55 2 0
1963 0.925 39165 34385 0.86 2.99 0.55 2 0
1968 0.904 39560 39560 1.28 4.28 0.85 2 0
1975 0.865 44952 43104 2.12 6.33 0.85 2 0
Dog and cat foods 1958 0.861 24539 20935 0.42 2.77 0.55 2 0
1963 0.943 38669 40458 0.59 3.88 0.85 2 0
1968 0.941 58155 58155 0.89 6.16 0.85 2 0
1975 0.944 74677 78807 1.34 9.66 0.85 2 0
Margarine 1958 0.881 50183 36611 1.46 7.58 0.55 2 0
1963 0.928 42188 36814 1.91 9.57 0.85 2 0
1968 0.938 31784 31784 2.89 10.59 0.85 2 0
Compound fat 1958 0.833 22975 15888 1.46 7.58 0.55 1 0
1963 0.848 21280 19094 1.91 9.57 0.55 1 0
1968 0.826 16498 16498 2.89 10.59 0.55 1 0
1975 0.884 27612 19640 3.78 13.51 0.55 1 0
Coffee, and coffee and 1958 0.979 22897 30630 0.63 3.82 0.55 2 0
chicory extracts and 1963 0.984 31801 34494 0.96 5.27 0.55 2 0
essences
1968 0.937 46642 46642 1.67 8.07 0.85 2 0
1975 0.927 51875 62435 2.51 9.78 0.85 2 0
Appendix B
Alcoholic cider, perry, apple 1958 0.812 20540 21770 0.67 3.46 0.55 2 0
pectin, and British wines 1963 0.972 20075 22784 1.20 4.86 0.55 2 0
1968 0.916 30004 30004 1.39 5.68 0.55 2 0
1975 0.902 46976 55645 2.75 7.55 0.55 2 0
Cigarettes 1963 0.997 1198822 1258907 4.46 4.84 0.55 2 0
Data Sets
1968 0.999 1236281 1236281 7.47 7.69 0.00 2 0
Other manufactured tobacco 1963 0.980 145500 158625 4.46 4.84 0.55 2 0
1968 0.981 129145 129145 7.47 7.69 0.25 2 0
1975 0.988 94571 119509 12.36 10.25 0.55 2 0
Lubricating oils and greases 1963 0.695 86247 82575 0.80 6.01 0.25 1 0
1968 0.716 93079 93079 1.28 10.93 0.25 1 0
1975 0.740 138734 135633 1.68 12.95 0.55 1 0
Pharmaceutical preparations 1958 0.286 119871 89003 1.33 4.10 0.55 2 0
1963 0.287 168815 138163 1.68 4.93 0.55 2 0
1968 0.349 221326 221326 2.67 7.55 0.55 2 0
1975 0.393 341185 474269 4.73 11.40 0.55 2 0
Hair preparations 1958 0.670 16860 18822 0.41 2.02 0.55 2 0
1963 0.542 26510 27535 0.48 2.03 0.55 2 0
1968 0.570 31731 31731 0.92 2.96 0.55 2 0
1975 0.607 46032 60769 1.50 5.06 0.55 2 0
Dental preparations 1963 0.944 13587 12347 0.48 2.03 0.85 2 0
1968 0.874 14450 14450 0.92 2.96 0.85 2 0
1975 0.803 21586 24625 1.50 5.06 0.85 2 0
Other toilet preparations 1958 0.420 33779 37708 0.41 2.02 0.25 2 0
1963 0.373 50961 52931 0.48 2.03 0.25 2 0
1968 0.405 70585 70585 0.92 2.96 0.25 2 0
1975 0.421 99401 111633 1.50 5.06 0.55 2 0
435
Table B2 (continued)
436
Industry Year C5 SS DS K=N K=L TVCLASS ADTYPE CHANGE
Soap 1958 0.832 59413 60862 2.27 7.35 0.55 2 0
1963 0.809 53272 53019 3.53 9.14 0.85 2 0
1968 0.801 48123 48123 4.44 14.43 0.85 2 0
1975 0.764 52683 49235 4.34 14.57 0.85 2 0
Finished detergents 1958 0.904 45542 36105 2.27 7.35 0.55 2 0
1963 0.845 55501 48629 3.53 9.14 0.85 2 0
1968 0.799 61265 61265 4.44 14.43 0.85 2 0
1975 0.828 90114 99083 4.34 14.57 0.85 2 0
Polishes 1958 0.671 17337 18133 0.36 2.20 0.25 2 0
1963 0.796 17901 18436 0.47 3.04 0.25 2 0
1968 0.809 17249 17249 0.85 4.29 0.55 2 0
1975 0.849 14138 18195 1.31 7.31 0.85 2 0
Pesticides, disinfectants, and 1958 0.424 29238 22616 0.41 2.96 0.55 2 0
household deodorisers 1963 0.520 34655 31491 0.68 4.56 0.55 2 0
1968 0.534 44697 44697 1.28 7.49 0.55 2 0
1975 0.533 73971 80699 2.57 13.13 0.55 2 0
Powered industrial trucks 1958 0.698 9424 10001 0.29 1.41 0.00 1 0
and industrial tractors 1963 0.764 16297 15212 0.33 1.56 0.00 1 0
1968 0.555 42182 42182 0.41 1.91 0.00 1 0
Appendix B
1975 0.607 71547 66984 0.53 2.44 0.00 1 0
Lawn mowers 1963 0.940 9840 9703 0.49 2.23 0.00 1 0
1968 0.882 18368 18368 0.59 2.82 0.00 1 0
1975 0.852 19401 19091 0.71 3.84 0.00 1 0
Photographic and cinemato- 1963 0.706 14917 15978 0.54 3.10 0.25 2 0
Data Sets
graphic apparatus and 1968 0.843 38812 38812 1.02 2.87 0.25 2 0
appliances, and document
1975 0.787 54097 63083 1.78 7.28 0.25 2 0
copying equipment
Television receiving sets 1958 0.528 78964 81098 0.53 1.10 0.00 1 0
1963 0.816 66324 72215 0.55 1.00 0.00 1 0
1968 0.927 86103 86103 1.01 1.98 0.00 1 0
1975 0.850 135769 201970 1.22 2.80 0.25 1 0
Radio receiving sets 1963 0.579 23118 25140 0.55 1.00 0.00 1 0
1968 0.747 11556 11556 1.01 1.98 0.00 1 0
1975 0.835 4939 7389 1.22 2.80 0.25 1 0
Cooking apparatus and 1958 0.651 20221 16031 1.18 2.55 0.00 2 1
appliances, electric 1963 0.600 26332 22655 1.18 2.27 0.25 2 1
1968 0.859 40547 40547 1.43 3.00 0.25 2 1
1975 0.791 41854 52480 2.14 4.34 0.25 2 1
Washing machines, 1958 0.764 34924 25410 1.18 2.55 0.25 2 1
electrically operated 1963 0.852 55575 44969 1.18 2.27 0.25 2 1
1968 0.869 36915 36915 1.43 3.00 0.25 2 1
1975 0.981 44509 52813 2.23 4.79 0.25 2 1
Batteries and accumulators 1958 0.782 42602 48030 0.49 1.72 0.25 1 1
1963 0.787 52821 56933 0.63 2.03 0.25 1 1
1968 0.841 66757 66757 0.98 2.68 0.55 1 1
1975 0.908 74751 76167 1.43 4.13 0.55 1 1
Electric lamps 1958 0.744 23666 18932 0.49 1.72 0.25 1 1
1963 0.717 35084 29795 0.63 2.03 0.25 1 1
437
1968 0.866 40308 40308 0.98 2.68 0.25 1 1
1975 0.941 35415 41413 1.43 4.13 0.25 1 1
Table B2 (continued)
438
Industry Year C5 SS DS K=N K=L TVCLASS ADTYPE CHANGE
Cars 1958 0.901 521809 471432 3.47 6.46 0.00 1 0
1963 0.912 752514 685258 3.49 5.81 0.00 1 0
1968 0.992 853240 853240 3.71 6.49 0.00 1 0
1975 0.982 690957 648622 3.78 6.78 0.00 1 0
Cutlery 1958 0.658 22897 20524 0.24 1.75 0.55 2 0
1963 0.669 29222 26752 0.39 2.20 0.55 2 0
1968 0.709 33214 33214 0.56 3.42 0.85 2 0
1975 0.560 25796 27019 0.60 4.41 0.85 2 0
Knitted, netted, or crocheted 1958 0.214 81626 55153 0.60 4.26 0.25 1 0
goods: hosiery 1963 0.201 82066 66593 0.68 4.23 0.25 1 0
1968 0.433 91573 91573 0.86 4.92 0.25 1 0
1975 0.455 58546 91096 0.99 5.81 0.25 1 0
Tufted carpets, carpeting, 1958 0.720 8129 7823 0.83 2.96 0.25 1 0
and carpet floor rugs 1963 0.507 24855 23456 0.91 3.16 0.25 1 0
1968 0.518 61323 61323 1.24 3.80 0.25 1 0
1975 0.450 101350 131887 1.93 6.23 0.25 1 0
Men’s and boys’ shirts, 1958 0.191 50765 50194 0.06 0.47 0.55 1 0
underwear, and nightwear 1963 0.254 54607 52636 0.08 0.65 0.25 1 0
1968 0.355 54724 54724 0.12 0.91 0.25 1 0
1975 0.400 51410 56144 0.14 1.11 0.25 1 0
Appendix B
Corsets and brassieres 1958 0.308 31864 31902 0.14 0.85 0.25 2 0
1963 0.380 40321 38997 0.13 0.91 0.00 2 0
1968 0.590 40537 40537 0.15 1.06 0.55 2 0
1975 0.632 32675 39991 0.17 1.32 0.85 2 0
Men’s and boys’ footwear 1958 0.222 65714 61826 0.11 0.67 0.25 1 0
Data Sets
(1958 S.I.C.) 1963 0.303 72032 67606 0.15 0.87 0.25 1 0
Women’s and girls’ 1958 0.239 90546 85189 0.11 0.67 0.25 1 0
footwear (1958 S.I.C.) 1963 0.289 98844 92770 0.15 0.87 0.25 1 0
Footwear, other than rubber 1963 0.242 183223 171964 0.15 0.87 0.25 1 0
footwear (1963 S.I.C.) 1968 0.316 175642 175642 0.20 1.15 0.55 1 0
1975 0.400 160203 180788 0.30 1.52 0.25 1 0
Domestic and fancy 1958 0.741 8502 8153 0.83 2.65 0.25 1 0
glassware 1963 0.721 11943 11346 1.41 3.68 0.00 1 0
1968 0.805 15646 15646 2.13 5.45 0.00 1 0
1975 0.777 20237 21335 2.89 8.70 0.00 1 0
Upholstered furniture 1958 0.186 50918 50660 0.10 0.78 0.00 1 0
1963 0.173 61289 61786 0.12 0.93 0.00 1 0
1968 0.173 70492 70492 0.19 1.38 0.00 1 0
1975 0.276 95909 102033 0.33 2.36 0.25 1 0
Domestic furniture, other 1963 0.160 91765 92509 0.12 0.93 0.00 1 0
than upholstered 1968 0.151 119446 119446 0.19 1.38 0.00 1 0
1975 0.217 184705 193848 0.33 2.36 0.25 1 0
Upholstered divan beds, 1958 0.399 32714 31871 0.19 1.38 0.00 1 0
mattresses, and other 1963 0.418 32665 32174 0.17 1.37 0.00 1 0
bedding 1968 0.451 38611 38611 0.15 1.32 0.55 1 0
1975 0.515 44476 41789 0.18 1.52 0.55 1 0
Published books 1958 0.311 62287 58950 0.32 2.48 0.00 1 0
1963 0.339 80119 78913 0.38 2.86 0.00 1 0
1968 0.322 106922 106922 0.49 3.85 0.00 1 0
439
1975 0.328 143567 134590 0.61 5.21 0.00 1 0
Table B2 (continued)
440
Industry Year C5 SS DS K=N K=L TVCLASS ADTYPE CHANGE
Tyres and tubes, other than 1958 0.932 112412 94782 1.27 3.80 0.25 1 1
retreaded tyres 1963 0.945 144054 130074 1.54 4.06 0.25 1 1
1968 0.928 184836 184836 2.39 5.90 0.55 1 1
1975 0.964 190550 199087 2.68 7.51 0.55 1 1
Brushes and brooms 1958 0.301 17362 19193 0.19 1.43 0.25 1 0
1963 0.380 18904 17959 0.26 1.78 0.55 1 0
1968 0.425 18763 18763 0.35 2.45 0.25 1 0
1975 0.455 18479 17809 0.42 3.39 0.85 1 0
Toys and indoor games 1958 0.469 37626 36656 0.19 1.17 0.25 2 0
1963 0.477 42598 41860 0.27 1.57 0.25 2 0
1968 0.582 65106 65106 0.44 2.10 0.55 2 0
1975 0.453 93046 85285 0.54 2.69 0.55 2 0
Notes: C5 is the five-firm sales concentration ratio.
SS is sales revenue (in £1000) deflated by the general producer price index.
DS is sales revenue (in £1000) deflated by industry-specific producer price indices.
K is the value of capital stock of the corresponding MLH industry in 1980 prices (in £ million).
L is employment of the corresponding MLH industry (in 1000).
N is the number of plants with at least 25 employees of the corresponding MLH industry.
TVCLASS is a measure of the importance of TV advertising in total media advertising (see text).
ADTYPE takes the value 1 for an industry with typical or average advertising-sales ratio between 1% and 2%, and the value 2 for an industry
Appendix B
with typical or average advertising-sales ratio higher than 2%.
CHANGE takes the value 1 for an industry with a change in competition regime and 0 otherwise.
Table B3
Data Sets
Data set for the advertising regressions of chapter 5
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Cigars 1954 0.062 1659 2.36 0.00 0.00 2 0
1958 0.108 2353 3.13 0.37 0.14 2 0
1963 0.228 4077 4.84 0.43 0.18 2 0
Cigarettes 1958 0.027 172726 3.13 0.39 0.14 2 0
1963 0.051 205131 4.84 0.45 0.18 2 0
Tobacco 1958 0.039 21991 3.13 0.24 0.14 2 0
1963 0.054 25720 4.84 0.38 0.18 2 0
Spirits, other than whisky and gin 1954 0.077 6591 2.64 0.00 0.00 2 0
1958 0.088 9233 2.80 0.15 0.06 2 0
1963 0.098 13271 4.45 0.13 0.08 2 0
1968 0.080 16827 6.24 0.19 0.09 2 0
1973 0.057 38074 9.42 0.26 0.09 2 0
Wines 1954 0.077 16507 4.05 0.00 0.00 2 0
1958 0.096 19474 3.46 0.28 0.14 2 0
1963 0.106 31933 4.86 0.23 0.18 2 0
1968 0.097 41033 5.68 0.43 0.20 2 0
1973 0.045 96514 7.02 0.49 0.20 2 0
Cider and perry 1954 0.052 5038 4.05 0.00 0.00 2 0
1958 0.119 10156 3.46 0.45 0.14 2 0
1963 0.121 6975 4.86 0.63 0.18 2 0
1968 0.113 10258 5.68 0.65 0.20 2 0
441
1973 0.057 15741 7.02 0.66 0.20 2 0
Table B3 (continued)
442
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Soft drinks 1954 0.017 54884 2.71 0.00 0.00 2 0
1958 0.028 76984 2.91 0.52 0.21 2 0
1963 0.025 84746 3.04 0.47 0.27 2 0
1968 0.017 109050 3.89 0.69 0.31 2 0
1973 0.017 188960 4.79 0.74 0.31 2 0
Margarine 1954 0.037 41207 5.65 0.00 0.00 2 0
1958 0.060 37644 7.58 0.42 0.21 2 0
1963 0.088 31576 9.57 0.75 0.27 2 0
1968 0.113 22486 10.59 0.84 0.31 2 0
1973 0.083 35703 14.72 0.88 0.31 2 0
Biscuits 1954 0.007 102953 1.75 0.00 0.00 2 1
1958 0.021 105596 2.63 0.47 0.21 2 1
1963 0.021 108920 3.69 0.58 0.27 2 1
1968 0.017 110530 4.30 0.91 0.31 2 1
1973 0.021 123231 5.27 0.98 0.31 2 1
Cereal foods 1963 0.106 31031 6.45 0.80 0.27 2 0
1968 0.104 35810 9.50 0.84 0.31 2 0
1973 0.084 42952 12.10 0.85 0.31 2 0
Canned meat and poultry 1954 0.001 75672 1.33 0.00 0.00 1 0
Appendix B
1958 0.008 73115 1.43 0.44 0.21 1 0
1963 0.017 75367 1.70 0.68 0.27 1 0
1968 0.013 97236 2.44 0.80 0.31 1 0
Canned fish 1954 0.008 17726 1.33 0.00 0.00 1 0
Data Sets
1958 0.011 40190 1.43 0.36 0.21 1 0
1963 0.012 30510 1.70 0.22 0.27 1 0
1968 0.006 41185 2.44 0.89 0.31 1 0
Meat and fish pastes 1954 0.039 5240 1.33 0.00 0.00 2 0
1958 0.041 6913 1.43 0.52 0.21 2 0
1963 0.015 7432 1.70 0.78 0.27 2 0
Canned vegetables 1954 0.013 51854 1.63 0.00 0.00 1 0
1958 0.015 61018 1.95 0.55 0.21 1 0
1963 0.017 68976 2.99 0.44 0.27 1 0
1968 0.017 78895 4.28 0.86 0.31 1 0
Frozen foods 1954 0.010 8693 1.48 0.00 0.00 2 0
1958 0.030 18688 1.67 0.71 0.21 2 0
1963 0.030 58854 2.26 0.74 0.27 2 0
1968 0.025 91278 3.23 0.83 0.31 2 0
Ice cream and ice lollies 1954 0.017 14817 3.72 0.00 0.00 2 0
1958 0.028 19234 3.65 0.63 0.14 2 0
1963 0.051 13726 4.74 0.77 0.18 2 0
1968 0.057 16717 5.81 0.72 0.20 2 0
1973 0.014 44027 7.34 0.57 0.20 2 0
Canned milk 1954 0.032 12178 3.72 0.00 0.00 2 1
1958 0.038 16747 3.65 0.35 0.21 2 1
1963 0.039 18249 4.74 0.35 0.27 2 1
1968 0.026 18820 5.81 0.81 0.31 2 1
1973 0.014 20198 7.34 0.99 0.31 2 1
443
Table B3 (continued)
444
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Potato crisps 1963 0.028 19905 2.99 0.94 0.27 2 0
1968 0.024 26886 4.28 0.87 0.31 2 0
Table jellies 1954 0.034 7042 1.63 0.00 0.00 2 0
1958 0.036 7170 1.95 0.83 0.06 2 0
1963 0.015 6411 2.99 0.62 0.08 2 0
1968 0.013 7101 4.28 0.12 0.09 2 0
Jams and marmalade 1954 0.010 38432 1.63 0.00 0.00 1 0
1958 0.010 39126 1.95 0.45 0.14 1 0
1963 0.027 38412 2.99 0.51 0.18 1 0
1968 0.016 34622 4.28 0.63 0.20 1 0
Sauces, pickles, and salad creams 1954 0.022 19178 1.63 0.00 0.00 2 0
1958 0.027 20709 1.95 0.40 0.14 2 0
1963 0.023 24460 2.99 0.21 0.18 2 0
1968 0.027 26367 4.28 0.42 0.20 2 0
Soups 1954 0.051 11652 1.63 0.00 0.00 2 0
1958 0.043 23580 1.95 0.49 0.21 2 0
1963 0.057 33115 2.99 0.45 0.27 2 0
1968 0.048 33233 4.28 0.76 0.31 2 0
Coffee and coffee extracts 1954 0.044 12625 2.88 0.00 0.00 2 0
Appendix B
1958 0.074 19922 3.82 0.48 0.21 2 0
1963 0.084 27602 5.27 0.63 0.27 2 0
1968 0.024 40452 8.07 0.74 0.31 2 0
1973 0.040 53582 9.64 0.84 0.31 2 0
Cocoa and drinking chocolate 1954 0.071 5796 1.95 0.00 0.00 2 1
Data Sets
1958 0.116 4625 2.42 0.40 0.21 2 1
1963 0.081 4442 3.40 0.63 0.27 2 1
1968 0.057 4415 4.64 0.90 0.31 2 1
1973 0.059 3652 5.90 1.00 0.31 2 1
Chocolate confectionery 1954 0.014 131442 1.95 0.00 0.00 2 1
1958 0.024 122200 2.42 0.57 0.21 2 1
1963 0.059 123018 3.40 0.77 0.27 2 1
1968 0.060 134432 4.64 0.98 0.31 2 1
1973 0.045 160991 5.90 0.95 0.31 2 1
Adhesive dressings 1954 0.081 2309 1.42 0.00 0.00 2 1
1958 0.077 2276 2.01 0.31 0.21 2 1
1963 0.069 3906 3.92 0.36 0.27 2 1
1968 0.039 5330 6.07 0.86 0.31 2 1
1973 0.021 4879 6.26 0.97 0.31 2 1
Sanitary towels 1954 0.019 8015 1.42 0.00 0.00 2 0
1958 0.013 9331 2.01 0.00 0.00 2 0
1963 0.027 10474 3.92 0.00 0.00 2 0
1968 0.031 11941 6.07 0.00 0.00 2 0
1973 0.060 11898 6.26 0.00 0.00 2 0
Dental preparations 1954 0.241 6445 1.98 0.00 0.00 2 0
1958 0.224 8703 2.02 0.54 0.21 2 0
1963 0.274 11123 2.03 0.87 0.27 2 0
1968 0.167 11713 2.96 0.81 0.31 2 0
1973 0.165 17081 4.27 0.85 0.31 2 0
445
Table B3 (continued)
446
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Hair preparations 1954 0.287 10239 1.98 0.00 0.00 2 0
1958 0.324 13267 2.02 0.50 0.14 2 0
1963 0.258 22605 2.03 0.40 0.18 2 0
1968 0.179 26631 2.96 0.47 0.20 2 0
1973 0.139 42914 4.27 0.63 0.20 2 0
Toilet soap 1954 0.105 12406 6.78 0.00 0.00 2 0
1958 0.155 15296 7.35 0.39 0.21 2 0
1963 0.190 15777 9.14 0.87 0.27 2 0
1968 0.139 15878 14.43 0.93 0.31 2 0
1973 0.066 24723 15.69 0.87 0.31 2 0
Other toilet preparations 1954 0.140 21117 1.98 0.00 0.00 2 0
1958 0.145 27282 2.02 0.28 0.06 2 0
1963 0.114 44477 2.03 0.26 0.08 2 0
1968 0.110 62266 2.96 0.27 0.09 2 0
1973 0.091 96613 4.27 0.51 0.09 2 0
Razors and blades 1954 0.071 5294 1.47 0.00 0.00 2 0
1958 0.087 6007 1.75 0.51 0.21 2 0
1963 0.115 6923 2.20 0.53 0.27 2 0
1968 0.222 7024 3.42 0.74 0.31 2 0
Appendix B
1973 0.117 7462 4.19 0.75 0.31 2 0
Toilet brushes 1954 0.060 2946 1.21 0.00 0.00 2 0
Data Sets
1958 0.090 2458 1.43 0.33 0.14 2 0
1963 0.039 3291 1.78 0.44 0.18 2 0
1968 0.027 3262 2.45 0.55 0.20 2 0
1973 0.020 4258 2.99 0.68 0.20 2 0
Paper handkerchiefs and facial tissues 1963 0.080 5449 4.84 0.45 0.18 2 0
1968 0.044 7833 6.08 0.84 0.20 2 0
1973 0.016 10665 7.90 0.43 0.20 2 0
Paper towels 1963 0.042 2670 4.84 0.03 0.18 2 0
1968 0.038 5455 6.08 0.66 0.20 2 0
1973 0.023 9012 7.90 0.85 0.20 2 0
Toilet paper 1963 0.039 12292 4.84 0.46 0.18 2 0
1968 0.009 17165 6.08 0.19 0.20 2 0
1973 0.008 28020 7.90 0.68 0.20 2 0
Domestic brushes and paintbrushes 1954 0.008 11114 1.21 0.00 0.00 1 0
1958 0.016 12261 1.43 0.27 0.06 1 0
1963 0.016 13675 1.78 0.49 0.08 1 0
1968 0.012 13171 2.45 0.10 0.09 1 0
1973 0.009 14023 2.99 0.28 0.09 1 0
Paint 1954 0.008 115533 2.64 0.00 0.00 2 1
1958 0.019 127330 3.19 0.24 0.14 2 1
1963 0.026 125338 3.64 0.39 0.18 2 1
1968 0.021 119802 4.37 0.56 0.20 2 1
1973 0.021 165041 5.52 0.63 0.20 2 1
447
Table B3 (continued)
448
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Soap and detergents 1954 0.083 59746 6.78 0.00 0.00 2 0
1958 0.138 65331 7.35 0.56 0.21 2 0
1963 0.132 67721 9.14 0.89 0.27 2 0
1968 0.095 73928 14.43 0.98 0.31 2 0
1973 0.067 92137 15.69 0.99 0.31 2 0
Cleansers 1954 0.268 3561 6.78 0.00 0.00 2 0
1958 0.337 6919 7.35 0.49 0.21 2 0
1963 0.337 9979 9.14 0.88 0.27 2 0
1968 0.333 8940 14.43 0.87 0.31 2 0
1973 0.167 11041 15.69 0.94 0.31 2 0
Floor and furniture polishes 1954 0.068 6596 1.69 0.00 0.00 2 0
1958 0.105 7478 2.20 0.45 0.21 2 0
1963 0.132 6927 3.04 0.43 0.27 2 0
1968 0.123 7791 4.29 0.75 0.31 2 0
1973 0.104 8024 5.95 0.96 0.31 2 0
Shoe polishes 1954 0.088 2508 1.69 0.00 0.00 2 0
1958 0.080 2864 2.20 0.23 0.06 2 0
1963 0.140 3186 3.04 0.23 0.08 2 0
1968 0.066 2909 4.29 0.57 0.09 2 0
Appendix B
1973 0.035 2512 5.95 0.01 0.09 2 0
Disinfectants and household deodorisers 1954 0.080 5768 2.54 0.00 0.00 2 0
Data Sets
1958 0.071 8564 2.96 0.54 0.21 2 0
1963 0.051 8194 4.56 0.58 0.27 2 0
1968 0.045 10840 7.49 0.80 0.31 2 0
1973 0.019 17517 14.92 0.81 0.31 2 0
Insecticides and pesticides 1954 0.024 7879 2.54 0.00 0.00 2 0
1958 0.026 11365 2.96 0.26 0.06 2 0
1963 0.034 11360 4.56 0.16 0.08 2 0
1968 0.029 12445 7.49 0.24 0.09 2 0
Dog and cat foods 1954 0.051 11133 2.38 0.00 0.00 2 0
1958 0.068 22702 2.77 0.53 0.21 2 0
1963 0.101 35482 3.88 0.86 0.27 2 0
1968 0.101 52577 6.16 0.98 0.31 2 0
1973 0.042 70972 8.81 0.95 0.31 2 0
Lawn mowers 1954 0.034 3030 1.87 0.00 0.00 2 0
1958 0.034 4448 2.07 0.03 0.00 2 0
1963 0.025 6908 2.23 0.00 0.00 2 0
1968 0.024 13963 2.82 0.03 0.00 2 0
1973 0.017 22309 3.65 0.04 0.00 2 0
Rubber gloves 1963 0.038 2469 0.73 0.00 0.27 2 0
1968 0.025 3337 0.92 0.98 0.31 2 0
1973 0.011 4424 1.20 0.99 0.31 2 0
Sewing machines 1958 0.137 2750 2.07 0.00 0.06 2 0
1963 0.117 2913 2.23 0.04 0.08 2 0
1968 0.094 3630 2.82 0.32 0.09 2 0
449
Table B3 (continued)
450
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Notepaper 1958 0.053 3122 1.76 0.02 0.00 2 1
1963 0.028 3907 2.58 0.19 0.00 2 1
1968 0.027 3050 3.36 0.00 0.00 2 1
Pens and mechanical pencils 1954 0.076 6633 1.77 0.00 0.00 2 0
1958 0.068 7969 1.63 0.29 0.06 2 0
1963 0.047 7925 2.33 0.00 0.08 2 0
1968 0.047 11423 3.41 0.18 0.09 2 0
1973 0.037 13506 5.18 0.26 0.09 2 0
Lead pencils 1954 0.009 1495 1.77 0.00 0.00 1 1
1958 0.018 1787 1.63 0.00 0.00 1 1
1963 0.014 1764 2.33 0.00 0.00 1 1
1968 0.006 1402 3.41 0.00 0.00 1 1
Toys and games 1963 0.015 40090 1.57 0.34 0.18 1 0
1968 0.020 59518 2.10 0.55 0.20 1 0
1973 0.016 95030 2.72 0.67 0.20 1 0
Baby carriages 1954 0.008 5616 0.94 0.00 0.00 1 0
1958 0.017 6052 1.17 0.47 0.00 1 0
1963 0.011 6965 1.57 0.11 0.00 1 0
1968 0.003 5963 2.10 0.00 0.00 1 0
Appendix B
1973 0.001 5771 2.72 0.00 0.00 1 0
Typewriters 1954 0.031 3978 2.83 0.00 0.00 2 1
Data Sets
1958 0.034 5091 2.64 0.03 0.00 2 1
1963 0.019 7315 3.14 0.01 0.00 2 1
1968 0.009 10096 4.48 0.00 0.00 2 1
Tape recorders 1958 0.038 3531 1.10 0.03 0.00 2 0
1963 0.024 9505 1.00 0.00 0.00 2 0
1968 0.023 6023 1.98 0.00 0.00 2 0
Radio and TV sets 1954 0.020 73585 0.97 0.00 0.00 1 0
1958 0.024 82120 1.10 0.04 0.00 1 0
1963 0.019 84998 1.00 0.04 0.00 1 0
1968 0.011 95841 1.98 0.05 0.00 1 0
1973 0.006 307292 2.32 0.13 0.00 1 0
Office furniture 1963 0.016 20428 1.87 0.00 0.00 1 0
1968 0.010 26308 2.24 0.00 0.00 1 0
Electric lamps and fluorescent lighting 1954 0.011 18503 1.64 0.00 0.00 1 1
1958 0.017 19963 1.72 0.29 0.06 1 1
1963 0.013 30653 2.03 0.12 0.08 1 1
1968 0.015 35515 2.68 0.32 0.09 1 1
1973 0.008 35630 3.71 0.04 0.09 1 1
Mattresses, divan beds, pillows 1954 0.008 45563 1.31 0.00 0.00 1 0
1958 0.008 45168 1.38 0.08 0.14 1 0
1963 0.010 37459 1.37 0.02 0.18 1 0
1968 0.011 47437 1.32 0.50 0.20 1 0
1973 0.010 59478 1.39 0.46 0.20 1 0
451
Table B3 (continued)
452
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Car batteries and accumulators 1954 0.016 22165 1.64 0.00 0.00 1 1
1958 0.012 21927 1.72 0.15 0.06 1 1
1963 0.011 26945 2.03 0.15 0.08 1 1
1968 0.009 34531 2.68 0.32 0.09 1 1
1973 0.002 45027 3.71 0.40 0.09 1 1
Tyres and tubes 1954 0.017 68407 2.88 0.00 0.00 1 1
1958 0.021 72141 3.80 0.17 0.14 1 1
1963 0.017 97444 4.06 0.22 0.18 1 1
1968 0.016 146660 5.90 0.40 0.20 1 1
1973 0.012 148535 7.87 0.47 0.20 1 1
Motorcycles, scooters, and mopeds 1954 0.012 12564 4.57 0.00 0.00 1 1
1958 0.023 16998 5.13 0.13 0.00 1 1
1963 0.023 10312 7.06 0.00 0.00 1 1
1968 0.014 7847 8.08 0.00 0.00 1 1
1973 0.015 15697 7.74 0.00 0.00 1 1
Cars 1954 0.011 180093 6.68 0.00 0.00 1 0
1958 0.012 255515 6.46 0.01 0.00 1 0
1963 0.012 415130 5.81 0.01 0.00 1 0
1968 0.010 537892 6.49 0.03 0.00 1 0
Appendix B
1973 0.012 906758 6.44 0.09 0.00 1 0
Industrial trucks and tractors 1963 0.021 10006 1.56 0.00 0.00 1 0
1968 0.014 30320 1.91 0.00 0.00 1 0
Men’s shirts, collars, and pyjamas 1954 0.006 46871 0.29 0.00 0.00 1 0
Data Sets
1958 0.014 48898 0.34 0.41 0.06 1 0
1963 0.014 54609 0.47 0.18 0.08 1 0
1968 0.012 55695 0.67 0.19 0.09 1 0
1973 0.004 69034 0.87 0.00 0.09 1 0
Women’s hosiery 1954 0.026 20947 3.76 0.00 0.00 1 0
1958 0.017 47937 4.26 0.09 0.06 1 0
1963 0.011 53552 4.23 0.09 0.08 1 0
1968 0.010 58079 4.92 0.18 0.09 1 0
1973 0.008 41551 5.91 0.30 0.09 1 0
Corsetry 1954 0.034 21044 0.88 0.00 0.00 2 0
1958 0.062 27406 0.85 0.18 0.14 2 0
1963 0.047 35272 0.91 0.03 0.18 2 0
1968 0.048 35927 1.06 0.46 0.20 2 0
1973 0.033 36739 1.21 0.84 0.20 2 0
Footwear 1954 0.009 199319 0.58 0.00 0.00 1 0
1958 0.010 184301 0.67 0.14 0.06 1 0
1963 0.010 214995 0.87 0.14 0.08 1 0
1968 0.011 214349 1.15 0.40 0.09 1 0
1973 0.006 269970 1.47 0.20 0.09 1 0
Domestic refrigerators and deep freezers 1958 0.037 14120 2.55 0.20 0.00 2 1
1963 0.032 16587 2.27 0.27 0.00 2 1
1968 0.011 22095 3.00 0.01 0.00 2 1
1973 0.022 31393 4.06 0.16 0.00 2 1
453
Table B3 (continued)
454
Industry Year ADS SSDOM K=L TVTOT TVINT ADTYPE CHANGE
Washing machines 1954 0.025 19213 2.19 0.00 0.00 2 1
1958 0.044 26214 2.55 0.22 0.06 2 1
1968 0.029 29250 3.00 0.33 0.09 2 1
1973 0.008 52081 4.06 0.26 0.09 2 1
Electric cookers and heaters 1954 0.016 18198 2.19 0.00 0.00 2 1
1958 0.033 24089 2.55 0.08 0.00 2 1
1963 0.042 44342 2.27 0.14 0.00 2 1
1968 0.018 52265 3.00 0.05 0.00 2 1
1973 0.009 61866 4.06 0.14 0.00 2 1
Appendix B
Table B4
Data Sets
Data set for the innovation regressions of chapter 6
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Lubricating oils and greases (MLH 1952–56 3 4 3 58551 63111 0.49 4.03 1 0
263) 1957–61 2 1 2 71533 74920 0.70 4.88 1 0
1962–66 0 0 0 74886 74886 0.80 6.01 1 0
1967–71 2 2 2 80197 83787 1.27 10.86 1 0
1972–76 0 0 0 103190 108339 1.60 11.97 1 0
General chemicals, organic, other 1952–56 0 0 0 98593 81489 5.00 21.25 2 0
than coal tar products (part of 1957–61 4 5 0 166999 143385 8.25 27.23 2 0
MLH 271.2)
1962–66 4 3 3 240323 240323 9.66 25.75 2 0
1967–71 3 4 3 312446 354349 16.29 49.23 2 0
1972–76 3 2 1 420695 551838 25.05 68.45 2 0
General chemicals, other than 1952–56 5 5 1 85966 71052 2.40 10.10 1 0
inorganic and organic (MLH 271.3) 1957–61 4 3 2 143114 122877 3.99 12.84 1 0
1962–66 4 4 2 168424 168424 4.67 17.77 1 0
1967–71 3 3 2 215759 244695 6.47 25.04 1 0
1972–76 5 9 2 280777 363533 7.59 26.70 1 0
Pharmaceutical chemicals and 1952–56 7 6 4 119569 81112 0.84 3.40 2 0
preparations (MLH 272) 1957–61 19 22 7 141742 120695 1.32 4.10 2 0
1962–66 15 9 7 194889 194889 1.68 4.93 2 0
1967–71 12 14 7 244569 300118 2.67 7.55 2 0
1972–76 12 16 8 363509 525543 4.46 11.30 2 0
455
Table B4 (continued)
456
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Paint (MLH 274) 1952–56 4 4 3 116353 109907 0.40 2.69 1 1
1957–61 6 8 2 132779 128358 0.57 3.19 1 1
1962–66 2 1 1 136354 136354 0.71 3.64 1 1
1967–71 4 4 4 128602 144445 0.92 4.37 1 1
1972–76 4 3 4 171899 193160 1.19 5.52 1 1
Soap and detergents (MLH 275) 1952–56 3 1 2 98982 104347 1.74 6.64 1 0
1957–61 3 3 0 111817 108659 2.27 7.41 1 0
1962–66 4 4 2 118241 118241 3.53 9.14 1 0
1967–71 3 3 1 128165 137219 4.44 14.43 1 0
1972–76 1 1 1 156682 173380 3.81 15.69 1 0
Synthetic resins and plastics 1952–56 9 8 1 109462 68260 4.66 13.02 2 0
materials and synthetic rubber 1957–61 11 14 3 171822 135849 5.12 12.18 2 0
(MLH 276) 1962–66 14 12 4 258679 258679 6.90 16.68 2 0
1967–71 8 8 5 366558 446994 7.90 19.17 2 0
1972–76 8 7 5 399964 610834 10.76 31.98 2 0
Dyestuffs and pigments (MLH 277) 1952–56 6 5 4 80988 77745 3.56 8.44 2 0
1957–61 11 11 5 72350 68237 4.79 9.73 2 0
1962–66 3 3 2 82818 82818 5.79 12.00 2 0
1967–71 11 11 7 89178 90339 10.71 20.56 2 0
Appendix B
1972–76 6 8 4 139149 157244 13.75 26.88 2 0
Formulated adhesives, gelatine, 1952–56 1 1 0 18435 16138 0.52 4.18 2 0
Data Sets
glue, size, etc. (MLH 279.2) 1957–61 0 0 0 19423 18319 0.70 5.57 2 0
1962–66 2 2 1 25493 25493 0.96 7.22 2 0
1967–71 3 3 2 32410 31232 1.39 7.99 2 0
1972–76 3 3 2 37579 36640 1.98 11.33 2 0
Formulated pesticides and 1952–56 0 0 0 20028 16106 0.37 2.27 2 0
disinfectants (MLH 279.4) 1957–61 2 3 2 25051 21331 0.40 2.72 2 0
1962–66 2 1 1 30600 30600 0.68 4.56 2 0
1967–71 2 2 2 38867 42791 1.28 7.49 2 0
1972–76 6 9 6 52194 72446 2.69 14.92 2 0
Surgical bandages, etc., and 1952–56 0 0 0 25749 19997 0.28 1.31 0 1
sanitary towels (MLH 279.6) 1957–61 1 2 1 23513 20804 0.55 2.01 0 1
1962–66 4 3 1 31660 31660 1.16 3.92 0 1
1967–71 2 2 0 37314 36288 2.12 6.07 0 1
1972–76 1 1 0 45700 42215 2.47 6.26 0 1
Aluminium and aluminium alloys 1952–56 8 6 8 143074 160803 1.05 3.23 0 1
(MLH 321) 1957–61 11 11 10 152853 158305 1.39 4.27 0 1
1962–66 7 6 6 192061 192061 1.55 5.02 0 1
1967–71 7 7 4 247753 232792 2.22 7.32 0 1
1972–76 5 4 5 279127 266303 4.66 15.66 0 1
Copper, brass and other copper 1952–56 0 0 0 301232 225263 0.92 4.44 0 1
alloys (MLH 322) 1957–61 0 1 0 241120 265403 1.26 5.22 0 1
1962–66 2 1 0 301753 301753 1.35 5.80 0 1
1967–71 2 3 2 420605 243113 1.64 6.71 0 1
1972–76 1 0 1 417250 220994 1.87 7.31 0 1
457
Table B4 (continued)
458
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Miscellaneous nonferrous metals 1952–56 4 4 4 140387 134577 1.56 6.21 0 0
(MLH 323) 1957–61 2 2 1 143488 150113 1.74 6.70 0 0
1962–66 2 2 1 174836 174836 1.83 7.17 0 0
1967–71 0 1 0 233154 178420 2.12 9.81 0 0
1972–76 2 2 0 207510 155632 2.63 14.65 0 0
Welding and flame-cutting 1952–56 1 2 1 12134 13683 0.65 3.06 1 0
equipment (part of MLH 332) 1957–61 5 4 4 18226 18627 0.80 3.54 1 0
1962–66 6 6 3 23137 23137 0.77 3.26 1 0
1967–71 2 2 0 30380 30802 0.96 4.02 1 0
1972–76 4 5 2 31123 29274 1.02 5.08 1 0
Industrial engines (MLH 334) 1952–56 0 0 0 62369 58111 2.03 3.86 1 1
1957–61 1 1 1 67277 56870 2.79 4.07 1 1
1962–66 0 0 0 85543 85543 3.02 4.11 1 1
1967–71 0 1 0 101817 96445 3.53 4.74 1 1
1972–76 3 3 3 104037 92112 5.10 6.08 1 1
Powered industrial trucks and 1952–56 0 0 0 8102 10043 0.22 1.13 1 0
industrial tractors (part of MLH 1957–61 0 0 0 11145 12673 0.29 1.41 1 0
337)
1962–66 2 2 2 17838 17838 0.33 1.56 1 0
1967–71 1 2 1 46045 49343 0.41 1.91 1 0
Appendix B
1972–76 2 1 1 70049 70136 0.57 2.44 1 0
Mining machinery (MLH 339.1) 1952–56 11 8 7 36875 41582 0.49 2.10 0 0
Data Sets
1957–61 10 14 8 53516 54693 0.50 2.09 0 0
1962–66 28 28 21 63789 63789 0.56 2.38 0 0
1967–71 24 25 18 71840 72838 0.70 2.81 0 0
1972–76 17 13 13 86122 82165 0.98 3.50 0 0
Printing, bookbinding, and paper 1952–56 3 3 3 29191 32917 0.30 1.91 1 0
goods making machinery (MLH 1957–61 2 3 2 25392 25951 0.38 2.15 1 0
339.2)
1962–66 8 14 5 35795 35795 0.49 2.43 1 0
1967–71 22 18 16 44586 45206 0.67 2.98 1 0
1972–76 13 10 7 50969 49067 0.92 4.10 1 0
Space-heating, ventilating, and air- 1952–56 4 6 4 31097 35067 0.26 1.76 0 0
conditioning equipment (MLH 1957–61 7 8 4 44873 45860 0.37 2.13 0 0
339.4) 1962–66 12 10 9 77232 77232 0.36 2.06 0 0
1967–71 4 3 1 95937 97270 0.41 2.40 0 0
1972–76 5 5 4 125158 120486 0.54 3.11 0 0
Packaging and bottling machinery 1952–56 1 1 1 7929 8942 0.30 1.85 0 0
(MLH 339.8) 1957–61 1 1 1 16473 16835 0.42 2.07 0 0
1962–66 2 5 2 19260 19260 0.49 2.23 0 0
1967–71 6 5 4 24115 24450 0.59 2.82 0 0
1972–76 4 4 4 30092 29197 0.65 3.65 0 0
Miscellaneous machinery, other 1952–56 11 13 8 93720 105683 0.30 1.85 0 0
than rolling mills (part of MLH 1957–61 17 15 12 113597 116097 0.42 2.07 0 0
339.9) 1962–66 19 21 9 138562 138562 0.49 2.23 0 0
1967–71 18 16 14 154218 156362 0.59 2.82 0 0
1972–76 12 17 11 228744 221406 0.65 3.65 0 0
459
Table B4 (continued)
460
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Constructional and other fabricated 1952–56 2 2 1 91924 103659 0.55 2.49 0 1
ironwork and steelwork (part of 1957–61 4 6 4 140853 143952 0.64 2.50 0 1
MLH 341) 1962–66 9 7 8 139156 139156 0.70 2.60 0 1
1967–71 1 1 1 170178 172544 0.77 2.80 0 1
1972–76 1 2 1 242387 233340 0.80 3.31 0 1
Photographic and document 1952–56 5 8 4 7762 6920 0.30 2.86 2 0
copying equipment (MLH 351) 1957–61 5 3 4 8318 7941 0.66 3.29 2 0
1962–66 3 2 2 12804 12804 0.54 3.10 2 0
1967–71 5 7 2 41956 39170 1.02 2.86 2 0
1972–76 8 6 2 57154 57442 1.51 5.64 2 0
Machinery for generating, 1952–56 7 9 6 82325 73401 0.91 1.64 2 1
transmitting and distributing 1957–61 7 5 7 99031 94553 1.12 1.87 2 1
electric power (part of MLH 361) 1962–66 4 7 4 120070 120070 1.10 1.98 2 1
1967–71 4 1 4 151267 157054 1.28 3.04 2 1
1972–76 3 3 3 145302 153358 1.35 3.76 2 1
Switchgear and controlling gear for 1952–56 1 1 1 80736 71984 0.91 1.64 2 1
motors (part of MLH 361) 1957–61 3 3 3 118969 113589 1.12 1.87 2 1
1962–66 2 3 2 139233 139233 1.10 1.98 2 1
1967–71 5 4 5 144897 150440 1.28 3.04 2 1
Appendix B
1972–76 2 2 2 115723 122139 1.35 3.76 2 1
Other electrical machinery (part of 1952–56 2 2 2 68706 61258 0.91 1.64 1 1
Data Sets
MLH 361) 1957–61 7 8 6 89776 85716 1.12 1.87 1 1
1962–66 6 4 4 89770 89770 1.10 1.98 1 1
1967–71 2 2 2 99909 103731 1.28 3.04 1 1
1972–76 1 2 1 99076 104570 1.35 3.76 1 1
Insulated wires and cables, other 1952–56 5 5 2 101215 90243 1.93 3.22 0 1
than telecommunication cables 1957–61 4 4 3 112680 107585 3.00 4.25 0 1
(part of MLH 362)
1962–66 1 1 1 148034 148034 2.84 4.68 0 1
1967–71 3 5 2 171854 152860 3.75 5.42 0 1
1972–76 6 4 6 188044 169877 5.12 7.57 0 1
Radio and electronic components 1952–56 19 23 6 74132 66096 0.44 1.62 2 0
(MLH 364) 1957–61 32 30 13 112946 107839 0.53 1.80 2 0
1962–66 19 19 7 151737 151737 0.60 1.63 2 0
1967–71 28 28 15 220987 229441 0.85 2.40 2 0
1972–76 23 30 16 353699 373310 1.36 3.35 2 0
Broadcast receiving and sound 1952–56 12 11 8 81867 72992 0.44 0.99 1 0
reproducing equipment, except 1957–61 10 11 8 101267 96688 0.53 1.10 1 0
gramophone records (part of MLH
1962–66 10 12 9 104905 104905 0.55 1.00 1 0
365)
1967–71 10 8 4 108476 112625 1.01 1.98 1 0
1972–76 2 2 1 262529 277085 1.39 2.32 1 0
Electronic computers (MLH 366) 1952–56 8 11 8 5305 4717 1.58 4.40 2 0
1957–61 8 9 4 7433 7097 1.84 4.83 2 0
1962–66 9 8 4 40141 40141 1.80 4.68 2 0
1967–71 7 6 5 100799 104655 2.30 4.21 2 0
1972–76 12 13 8 206781 218246 2.67 4.59 2 0
461
Table B4 (continued)
462
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Radio, radar and electronic capital 1952–56 7 9 6 50556 45075 1.30 2.80 2 0
goods (MLH 367) 1957–61 13 14 8 51668 49332 1.56 3.10 2 0
1962–66 21 20 17 127298 127298 1.37 2.80 2 0
1967–71 24 28 23 157694 163727 1.37 3.32 2 0
1972–76 28 31 26 218401 230510 1.65 3.64 2 0
Electric appliances primarily for 1952–56 1 1 1 89989 73899 0.64 2.11 1 1
domestic use (MLH 368) 1957–61 4 4 4 133477 118774 1.18 2.55 1 1
1962–66 0 1 0 200274 200274 1.18 2.27 1 1
1967–71 1 0 1 202937 227281 1.43 3.00 1 1
1972–76 1 1 0 231079 289800 2.18 4.06 1 1
Electrical equipment for motor 1952–56 2 2 1 60573 61414 0.41 1.64 1 0
vehicles, cycles, and aircraft (part 1957–61 3 3 2 71395 77665 0.49 1.68 1 0
of MLH 369) 1962–66 8 7 5 84342 84342 0.63 2.03 1 0
1967–71 3 3 2 108498 109530 0.98 2.68 1 0
1972–76 0 1 0 130750 134105 1.56 3.71 1 0
Primary batteries and accumulators 1952–56 0 0 0 38407 37471 0.41 1.64 1 1
(part of MLH 369) 1957–61 1 1 0 39663 41494 0.49 1.68 1 1
1962–66 0 0 0 49182 49182 0.63 2.03 1 1
1967–71 1 1 0 60498 56122 0.98 2.68 1 1
Appendix B
1972–76 1 1 1 71920 64716 1.56 3.71 1 1
Electric lamps, electric light 1952–56 2 1 2 58326 51798 0.41 1.64 1 1
Data Sets
fittings, and wiring accessories 1957–61 0 1 0 70465 64431 0.49 1.68 1 1
(part of MLH 369) 1962–66 5 5 3 97992 97992 0.63 2.03 1 1
1967–71 6 6 6 110277 116044 0.98 2.68 1 1
1972–76 3 4 3 136861 148468 1.56 3.71 1 1
Marine machinery (part of MLH 1952–56 5 3 4 109362 123322 0.43 0.98 1 0
370) 1957–61 3 5 1 78298 80021 0.59 1.25 1 0
1962–66 7 5 6 65119 65119 0.86 1.96 1 0
1967–71 2 3 2 71559 72553 0.98 2.11 1 0
1972–76 1 0 1 78174 76500 1.36 2.68 1 0
Motor vehicles and parts, except 1952–56 12 11 9 824394 899139 2.72 6.72 1 0
trailers and caravans (part of MLH 1957–61 16 18 13 1172390 1186438 3.47 6.46 1 0
381) 1962–66 14 14 8 1612064 1612064 3.49 5.81 1 0
1967–71 19 23 16 1820051 1869719 3.71 6.49 1 0
1972–76 21 23 19 2049130 2012628 4.01 6.43 1 0
Motorcycles (part of MLH 382) 1952–56 1 0 1 35061 31387 1.50 4.44 1 1
1957–61 1 1 1 27233 21586 1.81 5.13 1 1
1962–66 0 0 0 16224 16224 2.84 7.06 1 1
1967–71 1 1 1 18871 17139 3.65 8.08 1 1
1972–76 2 3 0 16263 14118 3.34 7.74 1 1
Paper and board (MLH 481) 1952–56 6 6 4 293269 269608 1.66 5.83 0 1
1957–61 5 8 5 321880 296653 2.74 8.74 0 1
1962–66 11 8 10 350074 350074 3.23 9.61 0 1
1967–71 5 6 3 377973 390695 4.91 13.74 0 1
1972–76 3 3 3 396899 408918 6.13 17.72 0 1
463
Table B4 (continued)
464
Industry Period INN1 INN2 INN3 SS DS K=N K=L RDTYPE CHANGE
Tyres and tubes (part of MLH 491) 1952–56 0 0 0 93731 99189 0.92 2.69 1 1
1957–61 1 1 0 96312 89924 1.26 3.80 1 1
1962–66 0 0 0 123440 123440 1.54 4.06 1 1
1967–71 1 3 1 158386 175366 2.39 5.90 1 1
1972–76 3 1 3 162691 182942 3.03 7.87 1 1
Plastic products (MLH 496) 1952–56 7 8 5 41734 35482 0.23 1.66 0 0
1957–61 6 6 4 65580 60985 0.26 1.86 0 0
1962–66 9 9 7 146583 146583 0.33 2.24 0 0
1967–71 18 24 15 249666 321945 0.55 3.60 0 0
1972–76 17 11 13 397786 574061 0.82 5.26 0 0
Notes: The classification according to MLH industries follows the 1968 Standard Industrial Classification.
INN1 is the number of innovations for time periods 1952–56, 1957–61, 1962–66, 1967–71, and 1972–76.
INN2 is the number of innovations for time periods 1953–57, 1958–62, 1963–67, 1968–72, and 1973–77.
INN3 is the number of innovations, excluding those reported as originating outside the UK, for time periods 1952–56, 1957–61, 1962–66, 1967–
71, and 1972–76.
SS is sales revenue (in £1000) deflated by the general producer price index for the years 1953, 1958, 1963, 1968, and 1973.
DS is sales revenue (in £1000) deflated by industry-specific producer price indices for the years 1953, 1958, 1963, 1968, and 1973.
K is the value of capital stock in 1980 prices (in £ million) of the corresponding MLH industry for the years 1953, 1958, 1963, 1968, and 1973.
L is employment (in 1000) of the corresponding MLH industry for the years 1953, 1958, 1963, 1968, and 1973.
N is the number of plants with at least 25 employees of the corresponding MLH industry for the years 1953, 1958, 1963, 1968, and 1973.
RDTYPE takes the value 0 for an industry with typical R&D-sales ratio lower than 1%, 1 for an industry with typical R&D-sales ratio between
Appendix B
1% and 2%, and 2 for an industry with typical R&D-sales ratio higher than 2%.
CHANGE takes the value 1 for an industry with a change in competition regime and 0 otherwise.
Table B5
Data Sets
Data set for the concentration regressions of chapter 6
Industry Year C5 SS DS K=N K=L RDTYPE CHANGE
Lubricating oils and greases 1963 0.695 86247 82575 0.80 6.01 1 0
1968 0.716 93079 93079 1.28 10.93 1 0
1975 0.740 138734 135633 1.68 12.95 1 0
Blacks, mineral and carbon 1963 0.984 10142 8622 7.33 27.33 1 0
1968 0.971 12250 12250 8.60 33.66 1 0
Organic chemicals: acids (carboxylic), acid 1963 0.632 16869 15955 9.66 25.75 1 0
anhydrides, and acid halides 1968 0.648 23035 23035 16.29 49.23 1 0
Organic chemicals: alcohols (monohydric) 1963 0.946 51182 43596 9.66 25.75 2 0
1968 0.980 48608 48608 16.29 49.23 2 0
Organic chemicals: esters 1963 0.864 11326 9036 9.66 25.75 1 0
1968 0.825 18389 18389 16.29 49.23 1 0
Organic chemicals: halogenated derivatives of 1963 0.974 29000 25566 9.66 25.75 2 0
hydrocarbons 1968 0.969 43843 43843 16.29 49.23 2 0
1975 0.999 57312 73666 23.40 70.00 2 0
Organic chemicals: aliphatic hydrocarbons 1963 0.774 26560 20123 9.66 25.75 2 0
1968 0.993 39191 39191 16.29 49.23 2 0
Additives for liquid fuels and lubricating oils 1963 0.901 30588 26966 4.67 17.77 1 0
1968 0.807 41886 41886 6.48 25.04 1 0
1975 0.904 60214 62292 7.49 25.59 1 0
Bleaching materials 1963 0.816 14602 15129 4.67 17.77 1 0
1968 0.840 15215 15215 6.48 25.04 1 0
465
1975 0.880 12900 15281 7.49 25.59 1 0
Table B5 (continued)
466
Industry Year C5 SS DS K=N K=L RDTYPE CHANGE
Chemicals and preparations mainly for PVC 1963 0.773 19818 17453 4.67 17.77 1 0
processing 1968 0.871 24061 24061 6.48 25.04 1 0
1975 0.750 32430 33732 7.49 25.59 1 0
Food and perfume materials 1963 0.710 20616 18854 4.67 17.77 1 0
1968 0.811 26590 26590 6.48 25.04 1 0
1975 0.768 31028 30866 7.49 25.59 1 0
Rubber-processing chemicals and preparations 1963 0.985 13418 11250 4.67 17.77 1 0
1968 0.940 17118 17118 6.48 25.04 1 0
1975 0.979 16830 16585 7.49 25.59 1 0
Surface active materials 1963 0.668 13815 13960 4.67 17.77 1 0
1968 0.551 21260 21260 6.48 25.04 1 0
1975 0.687 23122 24279 7.49 25.59 1 0
Pharmaceutical chemicals 1963 0.539 52708 42369 1.68 4.93 2 0
1968 0.610 60624 60624 2.67 7.55 2 0
1975 0.660 67767 67286 4.73 11.40 2 0
Pharmaceutical preparations 1958 0.286 119871 89003 1.33 4.10 2 0
1963 0.287 168815 138163 1.68 4.93 2 0
1968 0.349 221326 221326 2.67 7.55 2 0
1975 0.393 341185 474269 4.73 11.40 2 0
Appendix B
Hair preparations 1958 0.670 16860 18822 0.41 2.02 2 0
1963 0.542 26510 27535 0.48 2.03 2 0
1968 0.570 31731 31731 0.92 2.96 2 0
1975 0.607 46032 60769 1.50 5.06 2 0
Dental preparations 1963 0.944 13587 12347 0.48 2.03 2 0
Data Sets
1968 0.874 14450 14450 0.92 2.96 2 0
1975 0.803 21586 24625 1.50 5.06 2 0
Other toilet preparations 1958 0.420 33779 37708 0.41 2.02 2 0
1963 0.373 50961 52931 0.48 2.03 2 0
1968 0.405 70585 70585 0.92 2.96 2 0
1975 0.421 99401 111633 1.50 5.06 2 0
Soaps 1958 0.832 59413 60862 2.27 7.35 1 0
1963 0.809 53272 53019 3.53 9.14 1 0
1968 0.801 48123 48123 4.44 14.43 1 0
1975 0.764 52683 49235 4.34 14.57 1 0
Finished detergents 1958 0.904 45542 36105 2.27 7.35 1 0
1963 0.845 55501 48629 3.53 9.14 1 0
1968 0.799 61265 61265 4.44 14.43 1 0
1975 0.828 90114 99083 4.34 14.57 1 0
Synthetic resins, thermosetting 1963 0.528 27759 20436 6.90 16.68 2 0
1968 0.487 40393 40393 7.90 19.17 2 0
Plastics materials other than synthetic resins, 1963 0.787 27581 24240 6.90 16.68 2 0
thermosetting 1968 0.768 30576 30576 7.90 19.17 2 0
1975 0.794 27783 31102 8.22 28.64 2 0
Plastics materials other than synthetic resins, 1963 0.745 32007 25633 6.90 16.68 2 0
thermoplastic 1968 0.665 54709 54709 7.90 19.17 2 0
Synthetic rubber 1958 0.917 2841 1641 5.12 12.18 2 0
1963 0.994 27581 21292 6.90 16.68 2 0
1968 0.994 37863 37863 7.90 19.17 2 0
467
1975 0.919 47882 40493 8.22 28.64 2 0
Table B5 (continued)
468
Industry Year C5 SS DS K=N K=L RDTYPE CHANGE
Finished synthetic organic dyestuffs 1958 0.971 29134 26262 4.79 9.73 2 0
1963 0.937 36627 36242 5.79 12.00 2 0
1968 0.904 40970 40970 10.71 20.56 2 0
1975 0.877 52043 59336 12.16 30.25 2 0
Polishes 1958 0.671 17337 18133 0.36 2.20 1 0
1963 0.796 17901 18436 0.47 3.04 1 0
1968 0.809 17249 17249 0.85 4.29 1 0
1975 0.849 14138 18195 1.31 7.31 1 0
Gelatine, glue, size, and other adhesives 1963 0.488 28641 29712 0.96 7.22 2 0
1968 0.461 36415 36415 1.39 7.99 2 0
1975 0.442 48031 44083 1.86 11.61 2 0
Explosives and fireworks 1958 0.897 47722 46778 1.77 3.58 1 0
1963 0.864 38202 37976 2.84 5.62 1 0
1968 0.835 36778 36778 4.09 7.59 1 0
Pesticides, disinfectants, and household 1958 0.424 29238 22616 0.41 2.96 2 0
deodorisers 1963 0.520 34655 31491 0.68 4.56 2 0
1968 0.534 44697 44697 1.28 7.49 2 0
1975 0.533 73971 80699 2.57 13.13 2 0
Welding and flame-cutting machines 1963 0.638 11802 11638 0.77 3.26 1 0
Appendix B
1968 0.501 14590 14590 0.96 4.02 1 0
1975 0.630 18124 15254 0.87 4.85 1 0
Air and gas compressors and exhausters 1958 0.415 17780 17914 0.63 2.37 1 1
Data Sets
1963 0.481 24848 24502 0.71 2.72 1 1
1968 0.572 33439 33439 0.77 2.89 1 1
1975 0.719 36671 33365 1.03 3.98 1 1
Internal combustion engines, industrial 1958 0.740 44160 39449 2.79 4.07 1 1
1963 0.677 55144 58265 3.02 4.11 1 1
1975 0.889 85484 79010 6.04 7.22 1 1
Textile machinery 1958 0.647 51270 51656 0.53 2.42 1 0
1963 0.603 73907 72878 0.66 2.59 1 0
1968 0.592 101346 101346 0.72 2.97 1 0
1975 0.619 93273 88189 0.95 4.15 1 0
Powered industrial trucks and industrial tractors 1958 0.698 9424 10001 0.29 1.41 1 0
1963 0.764 16297 15212 0.33 1.56 1 0
1968 0.555 42182 42182 0.41 1.91 1 0
1975 0.607 71547 66984 0.53 2.44 1 0
Printing, bookbinding, paperworking, etc. 1958 0.446 20260 20413 0.38 2.15 1 0
machinery 1963 0.491 28556 28159 0.49 2.43 1 0
1968 0.491 33563 33563 0.67 2.98 1 0
1975 0.440 39662 37500 0.75 3.55 1 0
Rubber and plastics working machinery 1963 0.563 16686 16453 0.49 2.23 1 0
1968 0.506 24158 24158 0.59 2.82 1 0
1975 0.508 19540 19228 0.71 3.84 1 0
Photographic and cinematographic apparatus and 1963 0.706 14917 15978 0.54 3.10 2 0
appliances, and document copying equipment 1968 0.843 38812 38812 1.02 2.87 2 0
469
1975 0.787 54097 63083 1.78 7.28 2 0
Table B5 (continued)
470
Industry Year C5 SS DS K=N K=L RDTYPE CHANGE
Optical instruments and appliances 1963 0.507 9436 10108 0.36 1.35 2 0
1968 0.570 14143 14143 0.51 1.86 2 0
1975 0.572 20023 23349 0.68 3.11 2 0
Transformers for lighting, heating, and power 1958 0.499 43892 38349 1.12 1.87 2 1
1963 0.450 60208 55061 1.10 1.98 2 1
1968 0.767 55120 55120 1.28 3.04 2 1
1975 0.689 38395 38395 1.28 3.78 2 1
Starting and controlling gear for electric motors 1958 0.613 33487 29258 1.12 1.87 1 1
1963 0.479 39976 36558 1.10 1.98 1 1
1968 0.526 44871 44871 1.28 3.04 1 1
1975 0.556 46821 46821 1.28 3.78 1 1
Switchgear and switchboards 1958 0.616 79351 69329 1.12 1.87 2 1
1963 0.609 96849 88570 1.10 1.98 2 1
1968 0.774 100001 100001 1.28 3.04 2 1
1975 0.664 70307 70307 1.28 3.78 2 1
Line apparatus for long-distance communication 1958 0.939 11364 10457 2.04 2.08 2 1
1963 0.962 23632 22753 3.22 2.81 2 1
1968 1.000 40143 40143 2.77 3.28 2 1
1975 0.949 31233 34038 3.39 3.62 2 1
Appendix B
Passive components 1963 0.353 76422 73580 0.60 1.63 1 0
1968 0.335 125370 125370 0.85 2.40 1 0
1975 0.355 130312 133557 1.44 4.39 1 0
Television receiving sets 1958 0.528 78964 81098 0.53 1.10 1 0
Data Sets
1963 0.816 66324 72215 0.55 1.00 1 0
1968 0.927 86103 86103 1.01 1.98 1 0
1975 0.850 135769 201970 1.22 2.80 1 0
Radio receiving sets 1963 0.579 23118 25140 0.55 1.00 1 0
1968 0.747 11556 11556 1.01 1.98 1 0
1975 0.835 4939 7389 1.22 2.80 1 0
Gramophones, record players, and tape recorders 1958 0.680 29788 25884 0.53 1.10 1 0
1963 0.553 35794 34463 0.55 1.00 1 0
Electronic computers 1963 0.733 46845 45102 1.80 4.68 2 0
1968 0.872 117633 117633 2.30 4.22 2 0
1975 0.864 213692 232886 2.38 5.18 2 0
Radio communication equipment 1958 0.646 14442 13289 1.56 3.00 2 0
1963 0.638 38064 36648 1.37 2.80 2 0
1968 0.808 44742 44742 1.37 3.32 2 0
1975 0.894 61650 67188 1.55 3.75 2 0
Radar and electronic navigation aid equipment 1958 0.783 21413 19704 1.56 3.00 2 0
1963 0.722 64096 61712 1.37 2.80 2 0
1968 0.746 82918 82918 1.37 3.32 2 0
1975 0.807 86093 93826 1.55 3.75 2 0
Electronic measuring and testing instruments 1958 0.467 11790 10849 1.56 3.00 2 0
1963 0.369 18071 17399 1.37 2.80 2 0
Washing machines, electrically operated 1958 0.764 34924 25410 1.18 2.55 1 1
1963 0.852 55575 44969 1.18 2.27 1 1
1968 0.869 36915 36915 1.43 3.00 1 1
471
1975 0.981 44509 52813 2.23 4.79 1 1
Table B5 (continued)
472
Industry Year C5 SS DS K=N K=L RDTYPE CHANGE
Electrical equipment for motor vehicles, cycles, 1958 0.747 83330 89751 0.49 1.72 1 0
and aircraft, excluding accumulators 1963 0.701 98189 97269 0.63 2.03 1 0
1968 0.740 125913 125913 0.98 2.68 1 0
1975 0.807 123426 134512 1.43 4.13 1 0
Batteries and accumulators 1958 0.782 42602 48030 0.49 1.72 1 1
1963 0.787 52821 56933 0.63 2.03 1 1
1968 0.841 66757 66757 0.98 2.68 1 1
1975 0.908 74751 76167 1.43 4.13 1 1
Electric lamps 1958 0.744 23666 18932 0.49 1.72 2 1
1963 0.717 35084 29795 0.63 2.03 2 1
1968 0.866 40308 40308 0.98 2.68 2 1
1975 0.941 35415 41413 1.43 4.13 2 1
Marine machinery 1963 0.346 56700 49781 0.86 1.96 1 0
1968 0.574 60135 60135 0.98 2.11 1 0
1975 0.457 55276 56078 1.36 2.89 1 0
Cars 1958 0.901 521809 471432 3.47 6.46 1 0
1963 0.912 752514 685258 3.49 5.81 1 0
1968 0.992 853240 853240 3.71 6.49 1 0
1975 0.982 690957 648622 3.78 6.78 1 0
Appendix B
Internal combustion engines for motor vehicles 1963 0.931 80452 78340 3.49 5.81 1 0
1968 0.970 85532 85532 3.71 6.49 1 0
1975 0.914 48412 45773 3.78 6.78 1 0
Aero-engines 1958 0.940 152178 157306 2.24 2.32 2 0
Data Sets
1963 0.971 108523 107012 3.00 2.58 2 0
Cutlery 1958 0.658 22897 20524 0.24 1.75 1 0
1963 0.669 29222 26752 0.39 2.20 1 0
1968 0.709 33214 33214 0.56 3.42 1 0
1975 0.560 25796 27019 0.60 4.41 1 0
Tyres and tubes, other than retreaded tyres 1958 0.932 112412 94782 1.27 3.80 1 1
1963 0.945 144054 130074 1.54 4.06 1 1
1968 0.928 184836 184836 2.39 5.90 1 1
1975 0.964 190550 199087 2.68 7.51 1 1
Notes: C5 is the five-firm sales concentration ratio.
SS is sales revenue (in £1000) deflated by the general producer price index.
DS is sales revenue (in £1000) deflated by industry-specific producer price indices.
K is the value of capital stock of the corresponding MLH industry in 1980 prices (in £ million).
L is employment of the corresponding MLH industry (in 1000).
N is the number of plants with at least 25 employees of the corresponding MLH industry.
RDTYPE takes the value 1 for an industry with typical R&D-sales ratio between 1% and 2%, and 2 for an industry with typical R&D-sales ratio
higher than 2%.
CHANGE takes the value 1 for an industry with a change in competition regime and 0 otherwise.
473
Table B6
474
Data set for chapter 7
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Slate and slate products 1954 16 17 2601 498 0.192 0.15 0.64 0.455 0 0 1
1958 17 18 2249 477 0.212 0.13 0.69 0.408 0 0 1
1963 16 21 2084 467 0.224 0.12 1.17 0.464 0 0 1
1968 15 17 2169 528 0.243 0.21 2.35 0.395 0 0 1
Clay, brick earth, marl, shale, 1954 48 116 9927 3185 0.321 0.26 4.67 0.455 0 0 1
and chalk 1958 39 120 10955 3835 0.350 0.31 5.91 0.408 0 0 1
1963 24 87 16964 5632 0.332 0.63 7.74 0.464 0 0 1
1968 20 102 30359 15311 0.504 1.01 12.61 0.395 0 0 1
Iron ore and ironstone 1954 19 46 10093 3082 0.305 1.18 7.64 0.455 0 0 0
1958 19 44 11714 3976 0.339 1.55 9.78 0.408 0 0 0
1963 11 50 11162 4534 0.406 1.26 11.51 0.464 0 0 0
Salt 1954 8 23 14395 4177 0.290 1.36 6.31 0.455 0 0 0
1958 8 19 14449 5913 0.409 1.94 8.50 0.408 0 0 0
1963 6 18 18729 10628 0.567 3.17 15.14 0.464 0 0 0
Nonmetalliferous mining and 1954 18 45 8771 1589 0.181 0.41 3.39 0.455 0 0 0
quarrying other than salt 1958 20 40 7088 1015 0.143 0.45 4.38 0.408 0 0 0
1963 13 24 5197 1859 0.358 0.73 7.77 0.464 0 0 0
Grain milling (MLH 211) 1954 149 227 322879 30864 0.096 0.53 4.16 0.305 0 0 1
Appendix B
1958 134 223 306594 44288 0.144 0.65 4.82 0.283 0 0 1
1963 78 183 278566 46326 0.166 1.04 7.17 0.295 0 0 1
1968 61 149 283599 54094 0.191 1.50 10.38 0.346 0 0 1
1973 56 133 275183 44925 0.163 1.87 13.23 0.490 0 0 1
Bread and flour confectionery 1954 763 1096 241093 34421 0.143 0.32 3.17 0.305 0 0 1
Data Sets
(MLH 212) 1958 554 839 293119 45856 0.156 0.43 2.86 0.283 0 0 1
1963 271 728 364153 62594 0.172 0.54 2.58 0.295 0 0 1
1968 209 536 381206 71191 0.187 0.84 3.03 0.346 0 0 1
1973 250 507 421962 77066 0.183 0.98 3.30 0.490 0 0 1
Biscuits (MLH 213) 1954 77 111 115973 20859 0.180 0.88 1.76 0.305 2 0 1
1958 60 95 122312 21270 0.174 1.51 2.65 0.283 2 0 1
1963 37 75 133484 31649 0.237 2.28 3.74 0.295 2 0 1
1968 33 65 149081 35662 0.239 3.22 4.39 0.346 2 0 1
1973 29 54 149029 26332 0.177 4.42 5.37 0.490 2 0 1
Bacon and ham (cured and 1954 146 185 176828 10066 0.057 0.26 2.21 0.305 0 0 1
smoked), sausages and sausage 1958 134 168 146719 14883 0.101 0.33 2.49 0.283 0 0 1
meat 1963 90 138 157864 14325 0.091 0.50 2.08 0.295 0 0 1
1968 67 101 166037 18246 0.110 0.80 2.49 0.346 0 0 1
Preserved meat or fish, meat 1963 50 74 43787 11967 0.273 0.23 1.57 0.295 1 0 0
extracts, fish cured, smoked, or 1968 43 65 59556 13788 0.232 0.59 2.89 0.346 1 0 0
salted, etc.
Other meat products, offal, lard, 1963 103 173 67826 8308 0.122 0.17 2.22 0.295 0 0 0
etc. 1968 99 130 74425 11179 0.150 0.35 2.85 0.346 0 0 0
Butter 1954 27 46 47012 2728 0.058 0.99 13.15 0.305 0 0 1
1958 26 42 60707 3308 0.054 1.26 12.16 0.283 0 0 1
1963 15 33 59916 2701 0.045 1.61 15.83 0.295 0 0 1
1968 17 38 74121 4327 0.058 1.59 16.08 0.346 0 0 1
475
Table B6 (continued)
476
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Condensed milk 1954 9 16 24685 1553 0.063 1.67 9.39 0.305 2 0 1
1958 7 19 36241 2871 0.079 1.73 8.85 0.283 2 0 1
1963 5 14 31741 5205 0.164 3.37 16.58 0.295 2 0 1
1968 6 10 29158 4758 0.163 5.86 21.74 0.346 2 0 1
Milk powder 1954 9 15 15664 913 0.058 0.73 7.54 0.305 0 0 1
1958 5 9 12789 648 0.051 0.94 8.78 0.283 0 0 1
1963 3 5 7801 887 0.114 2.42 16.35 0.295 0 0 1
Cream and other milk products 1954 14 22 20159 2894 0.144 0.97 6.62 0.305 1 0 1
(1958 S.I.C.) 1958 12 21 19366 3469 0.179 1.25 8.28 0.283 1 0 1
1963 15 27 27786 3516 0.127 1.56 9.19 0.295 1 0 1
Ice cream 1954 22 28 20749 5140 0.248 1.65 8.97 0.305 2 0 0
1958 14 22 28054 8143 0.290 4.18 12.27 0.283 2 0 0
1963 13 22 19029 4039 0.212 3.97 20.86 0.295 2 0 0
Cocoa and chocolate products 1954 58 79 185400 26256 0.142 1.28 2.01 0.305 2 0 1
1958 53 75 174518 32179 0.184 2.10 2.85 0.283 2 0 1
1963 41 65 175872 43085 0.245 3.63 4.20 0.295 2 0 1
1968 30 50 195358 43690 0.224 5.79 5.88 0.346 2 0 1
Fruit and vegetable products 1954 209 309 152109 21938 0.144 0.29 1.65 0.305 2 0 0
(MLH 218) 1958 168 263 190250 30516 0.160 0.44 1.97 0.283 2 0 0
Appendix B
1963 131 223 248688 51160 0.206 0.88 3.13 0.295 2 0 0
1968 119 201 286069 60947 0.213 1.29 4.41 0.346 2 0 0
1973 110 172 369177 76929 0.208 1.96 5.34 0.490 2 0 0
Dog and cat foods 1954 10 12 9937 2148 0.216 0.43 2.53 0.305 2 0 0
Data Sets
1958 12 14 20516 4363 0.213 0.58 2.72 0.283 2 0 0
1963 8 15 30693 9976 0.325 1.05 4.71 0.295 2 0 0
1968 12 17 44694 9639 0.216 1.67 6.61 0.346 2 0 0
Vegetable and animal oils and 1954 47 74 131158 12767 0.097 0.74 5.15 0.330 0 0 0
fats (MLH 221) 1958 44 76 158533 11888 0.075 0.89 6.15 0.311 0 0 0
1963 45 76 127451 13258 0.104 1.02 8.43 0.334 0 0 0
1968 45 77 113298 12776 0.113 1.24 12.30 0.394 0 0 0
1973 41 65 173956 23076 0.133 1.95 16.13 0.471 0 0 0
Margarine (MLH 229.1) 1954 17 18 68701 6919 0.101 1.73 5.85 0.305 2 0 0
1958 12 14 62065 6364 0.103 2.81 7.83 0.283 2 0 0
1963 8 9 52338 3932 0.075 4.47 9.86 0.295 2 0 0
1968 7 8 40621 2240 0.055 5.43 10.76 0.346 2 0 0
1973 7 8 59649 5270 0.088 7.18 14.96 0.490 2 0 0
Coffee, and coffee and chicory 1954 7 9 11152 1192 0.107 0.47 4.45 0.305 2 0 0
extracts and essences 1958 7 10 21788 3821 0.175 1.16 6.30 0.283 2 0 0
1963 9 13 29465 6999 0.238 1.58 6.44 0.295 2 0 0
Spirits, distilled 1958 16 65 85846 4758 0.055 0.34 6.20 0.283 0 0 1
1963 16 70 72550 6366 0.088 0.67 11.62 0.295 0 0 1
1968 17 74 118681 9181 0.077 0.89 15.39 0.346 0 0 1
British wines, cider, and perry 1954 18 30 13693 3312 0.242 0.54 4.78 0.305 2 0 0
(MLH 239.2) 1958 18 26 18830 6327 0.336 0.67 4.15 0.283 2 0 0
1963 7 17 19466 6030 0.310 1.20 5.21 0.295 2 0 0
1968 10 19 27750 6841 0.247 1.32 5.85 0.346 2 0 0
1973 9 17 41139 13033 0.317 2.02 7.23 0.490 2 0 0
477
Table B6 (continued)
478
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Tobacco (MLH 240) 1958 26 62 1053944 65865 0.062 2.24 3.14 0.570 2 0 0
1963 16 47 1224000 79042 0.065 4.46 4.86 0.673 2 0 0
1968 12 41 1255888 100639 0.080 7.66 7.73 0.769 2 0 0
1973 10 34 1134385 117867 0.104 11.65 10.10 0.994 2 0 0
Coke and manufactured fuel 1954 28 92 166172 15091 0.091 3.24 14.14 0.330 0 0 0
(MLH 261) 1958 25 83 214270 22461 0.105 6.02 21.60 0.311 0 0 0
1963 24 56 198204 12943 0.065 9.21 28.43 0.334 0 0 0
Lubricating oils and greases 1954 54 82 55148 13252 0.240 0.46 5.15 0.330 1 1 0
(MLH 263) 1958 46 73 59879 15703 0.262 0.62 5.92 0.311 1 1 0
1963 40 67 57484 16006 0.278 0.72 7.31 0.334 1 1 0
1968 37 61 58883 13796 0.234 1.25 12.75 0.394 1 1 0
1973 34 54 83814 17366 0.207 1.57 14.06 0.471 1 1 0
Miscellaneous basic chemicals 1958 96 182 174748 60730 0.348 5.51 23.47 0.311 0 1 0
(1958 S.I.C.) 1963 104 207 263477 99773 0.379 6.44 32.45 0.334 0 1 0
General chemicals, other than 1963 97 141 171835 57316 0.334 4.11 18.87 0.334 0 1 0
inorganic and organic (MLH 1968 88 127 211793 62688 0.296 6.27 27.19 0.394 0 1 0
271.3, 1968 S.I.C.) 1973 91 132 341297 83509 0.245 7.36 28.99 0.471 0 1 0
Pharmaceutical chemicals and 1954 152 184 136436 45331 0.332 1.00 4.25 0.330 2 2 0
preparations (MLH 272) 1958 130 160 169296 57592 0.340 1.45 4.57 0.311 2 2 0
Appendix B
1963 129 182 227398 89222 0.392 1.64 5.09 0.334 2 2 0
1968 103 150 279096 111350 0.399 2.81 7.92 0.394 2 2 0
1973 117 148 414171 153658 0.371 4.70 11.86 0.471 2 2 0
Toilet preparations (MLH 273) 1954 57 60 37424 15031 0.402 0.37 2.03 0.330 2 2 0
Data Sets
1958 59 64 48508 18264 0.377 0.41 2.09 0.311 2 2 0
1963 69 77 73365 31103 0.424 0.50 2.19 0.334 2 2 0
1968 56 67 100269 38634 0.385 1.03 3.17 0.394 2 2 0
1973 69 69 137667 54393 0.395 1.47 4.58 0.471 2 2 0
Paint (MLH 274) 1954 198 272 128395 29220 0.228 0.39 3.04 0.330 2 1 1
1958 173 248 141652 32847 0.232 0.53 3.47 0.311 2 1 1
1963 136 219 148773 38306 0.257 0.64 3.93 0.334 2 1 1
1968 96 146 143872 39135 0.272 0.97 4.81 0.394 2 1 1
1973 99 131 173887 54041 0.311 1.25 6.07 0.471 2 1 1
Soap and detergents (MLH 275) 1954 66 82 109731 20591 0.188 1.84 7.24 0.330 2 1 0
1958 56 73 125164 27885 0.223 2.27 7.91 0.311 2 1 0
1963 42 55 134478 35697 0.265 3.66 9.78 0.334 2 1 0
1968 37 48 128781 32399 0.252 4.81 16.03 0.394 2 1 0
1973 45 58 146130 32968 0.226 4.14 17.43 0.471 2 1 0
Synthetic resins and plastics 1954 — 63 94210 21470 0.228 4.29 12.88 0.330 0 2 0
materials 1958 57 78 146874 28741 0.196 4.38 11.56 0.311 0 2 0
1963 62 101 194080 46757 0.241 5.08 15.96 0.334 0 2 0
1968 85 128 324191 82940 0.256 6.92 19.15 0.394 0 2 0
Synthetic rubber 1963 4 4 22560 6632 0.294 27.28 72.01 0.334 0 2 0
1968 4 5 34617 10825 0.313 24.18 62.64 0.394 0 2 0
Dyestuffs 1954 18 28 73182 15085 0.206 4.21 6.13 0.330 0 2 0
1958 15 23 58603 13441 0.229 5.10 6.77 0.311 0 2 0
1963 15 27 71993 18408 0.256 6.56 10.52 0.334 0 2 0
1968 15 20 79126 22532 0.285 14.04 17.77 0.394 0 2 0
479
Table B6 (continued)
480
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Polishes (MLH 279.1) 1954 38 50 19930 4925 0.247 0.20 1.93 0.330 2 1 0
1958 30 38 21543 6875 0.319 0.35 2.51 0.311 2 1 0
1963 34 46 26856 7891 0.294 0.46 3.35 0.334 2 1 0
1968 29 37 33787 10921 0.323 0.80 4.70 0.394 2 1 0
1973 29 33 38620 12772 0.331 1.10 6.52 0.471 2 1 0
Formulated adhesives, gelatine, 1963 36 55 26219 5894 0.225 0.84 7.86 0.334 0 2 0
glue, size, etc. (MLH 279.2) 1968 29 42 39693 9217 0.232 1.39 9.00 0.394 0 2 0
1973 25 36 44826 11298 0.252 1.98 12.78 0.471 0 2 0
Fireworks 1954 17 21 3280 989 0.302 0.04 0.32 0.330 0 0 1
1958 15 20 3646 857 0.235 0.06 0.41 0.311 0 0 1
1963 12 20 3554 1029 0.290 0.11 0.89 0.334 0 0 1
Explosives 1954 9 42 65707 12172 0.185 1.59 1.60 0.330 0 1 0
1958 9 37 51972 12527 0.241 2.98 3.94 0.311 0 1 0
1963 9 20 41663 12129 0.291 5.63 6.28 0.334 0 1 0
1968 6 12 39978 11432 0.286 10.30 8.25 0.394 0 1 0
Formulated pesticides and 1954 35 42 18211 3791 0.208 0.31 2.56 0.330 2 2 0
disinfectants (MLH 279.4) 1958 39 49 22835 5787 0.253 0.34 2.96 0.311 2 2 0
1963 30 40 27686 8439 0.305 0.62 4.83 0.334 2 2 0
1968 25 33 32399 7832 0.242 1.20 8.14 0.394 2 2 0
Appendix B
1973 22 22 65472 19064 0.291 2.44 16.21 0.471 2 2 0
Surgical bandages, etc. and 1963 22 42 35148 11317 0.322 1.13 4.04 0.562 0 0 1
sanitary towels (MLH 279.6) 1968 18 37 42561 14561 0.342 2.00 6.11 0.602 0 0 1
1973 16 34 53802 16919 0.314 2.32 6.30 0.722 0 0 1
Steel forgings 1958 23 29 13645 2756 0.202 0.52 3.95 0.499 0 0 1
Data Sets
1963 26 35 24829 3362 0.135 0.92 7.05 0.528 0 0 1
Steel sheets and tinplate 1958 17 34 174834 11601 0.066 4.97 7.54 0.499 0 0 1
1963 11 19 105977 8763 0.083 7.08 12.96 0.528 0 0 1
Steel manufacture without 1958 130 184 259280 21407 0.083 0.90 4.13 0.499 0 0 0
melting, other than forgings, 1963 99 154 294025 25729 0.088 2.29 7.28 0.528 0 0 0
sheets, and tinplate
Steel castings 1958 52 63 45413 9411 0.207 0.84 2.42 0.499 0 0 1
1963 51 66 47703 9490 0.199 1.19 3.75 0.528 0 0 1
Steel manufacture with melting, 1958 48 95 897286 142228 0.159 28.73 14.82 0.499 0 0 0
other than castings 1963 41 97 822988 127014 0.154 41.75 22.42 0.528 0 0 0
Steel tubing and gas cylinders 1958 22 55 189335 33744 0.178 4.53 5.67 0.499 0 0 1
1963 23 61 187372 28462 0.152 6.25 8.44 0.528 0 0 1
1968 30 70 209559 39712 0.190 6.09 9.10 0.567 0 0 1
Steel tubes, manipulated, 1958 42 63 23553 4069 0.173 0.25 1.81 0.499 0 0 1
fabricated, etc. 1963 52 69 31605 6555 0.207 0.38 2.71 0.528 0 0 1
1968 39 53 34377 7230 0.210 0.71 3.81 0.567 0 0 1
Pig iron 1958 24 24 55198 5607 0.102 2.28 7.12 0.499 0 0 0
1963 10 13 37820 4241 0.112 4.52 12.46 0.528 0 0 0
Pressure pipes and fittings 1958 21 32 39405 5175 0.131 3.37 7.90 0.499 0 0 1
1963 15 25 38602 5820 0.151 4.02 8.23 0.528 0 0 1
1968 11 23 38671 3970 0.103 3.46 5.90 0.567 0 0 1
Iron castings for the building 1958 53 71 32110 6252 0.195 0.20 0.89 0.499 0 0 1
industry 1963 39 57 37901 9914 0.262 0.51 2.02 0.528 0 0 1
481
1968 29 42 35835 8025 0.224 1.27 4.22 0.567 0 0 1
Table B6 (continued)
482
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Ingot moulds and bottoms 1958 19 19 16357 2386 0.146 2.17 10.37 0.499 0 0 1
1963 13 14 15561 3217 0.207 2.64 10.38 0.528 0 0 1
1968 9 17 18545 3370 0.182 1.86 7.74 0.567 0 0 1
Marine and other engineering 1958 321 341 58702 8119 0.138 0.17 1.71 0.499 0 0 1
castings 1963 258 278 54862 9124 0.166 0.32 3.12 0.528 0 0 1
1968 183 208 48501 9335 0.192 0.49 4.57 0.567 0 0 1
Aluminium and aluminium 1954 23 36 98896 10417 0.105 3.64 4.39 0.509 0 0 1
alloys: fabricated 1958 27 44 106869 12093 0.113 4.00 5.46 0.499 0 0 1
1963 35 63 145189 19994 0.138 3.68 6.44 0.528 0 0 1
1968 38 62 171511 20356 0.119 4.95 9.26 0.567 0 0 1
Aluminium and aluminium 1954 76 91 30411 3534 0.116 0.28 1.91 0.509 0 0 1
alloys: founded 1958 74 91 31342 4292 0.137 0.35 2.39 0.499 0 0 1
1963 90 115 41397 7470 0.180 0.46 3.07 0.528 0 0 1
1968 90 120 57766 9736 0.169 0.89 5.17 0.567 0 0 1
Copper, brass and other copper 1954 24 28 64971 2072 0.032 0.56 4.92 0.509 0 0 1
alloys: smelted and refined 1958 23 27 42234 2209 0.052 0.51 5.90 0.499 0 0 1
1963 31 42 52461 3415 0.065 0.36 3.59 0.528 0 0 1
1968 20 26 105851 5909 0.056 1.36 10.28 0.567 0 0 1
Copper, brass and other copper 1954 169 207 227902 24357 0.107 1.03 4.82 0.509 0 0 1
Appendix B
alloys: fabricated and founded 1958 136 179 202096 23157 0.115 1.35 5.47 0.499 0 0 1
1963 126 176 259488 33028 0.127 1.55 6.41 0.528 0 0 1
1968 116 168 342383 37919 0.111 1.73 7.06 0.567 0 0 1
Miscellaneous nonferrous metals 1954 89 123 165314 16254 0.098 1.40 7.54 0.509 0 0 0
Data Sets
(MLH 323) 1958 97 127 175503 16495 0.094 1.52 7.04 0.499 0 0 0
1963 79 123 199829 22709 0.114 1.74 7.59 0.528 0 0 0
1968 80 115 310325 32855 0.106 2.11 10.34 0.567 0 0 0
1973 78 107 283136 42038 0.148 2.60 15.45 0.681 0 0 0
Welding and flame-cutting 1954 29 35 17504 2657 0.152 0.74 4.38 0.509 0 1 0
equipment 1958 31 40 23903 4203 0.176 0.76 4.04 0.499 0 1 0
1963 31 36 18480 3443 0.186 0.60 3.49 0.528 0 1 0
1968 33 44 32371 6141 0.190 0.48 2.65 0.567 0 1 0
Industrial engines (MLH 334) 1954 41 58 70673 13542 0.192 2.24 4.00 0.509 0 1 1
1958 35 55 80071 12064 0.151 2.89 4.17 0.499 0 1 1
1963 27 50 94382 13903 0.147 3.08 4.17 0.528 0 1 1
1968 21 44 115623 23716 0.205 3.61 4.77 0.567 0 1 1
1973 16 30 107459 20053 0.187 5.27 6.12 0.681 0 1 1
Textile machinery and 1954 266 338 95689 21590 0.226 0.36 1.99 0.509 0 1 0
accessories (MLH 335) 1958 189 245 77793 14165 0.182 0.52 2.69 0.499 0 1 0
1963 148 204 113738 28142 0.247 0.65 2.83 0.528 0 1 0
1968 146 213 140642 35674 0.254 0.69 3.22 0.567 0 1 0
1973 123 176 143124 39012 0.273 0.91 4.31 0.681 0 1 0
Lifts and escalators 1954 32 54 10618 2835 0.267 0.04 0.36 0.509 0 0 1
1958 30 62 15773 3688 0.234 0.05 0.39 0.499 0 0 1
1963 32 70 29641 7099 0.239 0.18 1.10 0.528 0 0 1
1968 28 62 38827 11528 0.297 0.33 1.65 0.567 0 0 1
483
Table B6 (continued)
484
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Powered industrial trucks and 1954 7 9 3884 211 0.054 0.19 1.27 0.509 1 1 0
industrial tractors 1958 6 6 7055 981 0.139 0.32 0.86 0.499 1 1 0
1963 17 23 19281 3175 0.165 0.22 0.92 0.528 1 1 0
1968 28 35 36516 9248 0.253 0.51 2.15 0.567 1 1 0
Mining machinery (MLH 339.1) 1954 53 66 41508 8456 0.204 0.50 2.13 0.509 0 0 0
1958 67 80 60392 11032 0.183 0.52 2.12 0.499 0 0 0
1963 66 84 68058 15533 0.228 0.57 2.44 0.528 0 0 0
1968 63 91 85380 16290 0.191 0.70 2.91 0.567 0 0 0
1973 56 75 99856 22153 0.222 0.97 3.64 0.681 0 0 0
Printing, bookbinding, and 1954 81 100 26214 7275 0.278 0.32 1.96 0.509 0 1 0
paper goods making machinery 1958 71 86 26670 7129 0.267 0.39 2.15 0.499 0 1 0
(MLH 339.2) 1963 66 93 35247 10502 0.298 0.48 2.59 0.528 0 1 0
1968 71 103 54335 12159 0.224 0.63 3.18 0.567 0 1 0
1973 72 79 62216 12899 0.207 0.87 4.37 0.681 0 1 0
Space-heating, ventilating and 1954 125 148 30226 6520 0.216 0.13 1.15 0.509 0 0 0
air-conditioning equipment 1958 121 136 39389 7976 0.203 0.15 1.11 0.499 0 0 0
(MLH 339.4)
1963 129 167 64507 16584 0.257 0.25 1.75 0.528 0 0 0
1968 171 221 96991 24068 0.248 0.38 2.80 0.567 0 0 0
Scales and weighing machinery 1954 17 45 9621 2682 0.279 0.23 1.88 0.509 0 0 0
Appendix B
1958 15 44 8447 2010 0.238 0.22 1.77 0.499 0 0 0
1963 14 50 11828 3291 0.278 0.23 1.93 0.528 0 0 0
1968 18 29 14481 4293 0.296 0.42 2.24 0.567 0 0 0
Pulpmaking and papermaking 1954 11 14 7933 1701 0.214 0.61 2.93 0.509 0 0 0
Data Sets
machinery 1958 16 19 17846 3427 0.192 0.73 2.59 0.499 0 0 0
1963 11 17 24607 4360 0.177 1.25 4.59 0.528 0 0 0
Packaging and bottling 1963 37 52 19727 4576 0.232 0.27 1.31 0.528 0 0 0
machinery 1968 47 63 29047 7010 0.241 0.67 3.58 0.567 0 0 0
Constructional and other 1963 292 359 139771 26769 0.192 0.37 2.70 0.528 0 0 1
fabricated ironwork and 1968 406 493 193201 39907 0.207 0.33 2.66 0.567 0 0 1
steelwork
Ordnance and small arms (MLH 1958 42 75 69364 11952 0.172 1.50 2.78 0.499 0 0 0
342) 1963 32 46 54281 10584 0.195 2.36 3.49 0.528 0 0 0
1968 30 40 54619 13273 0.243 2.92 5.16 0.567 0 0 0
1973 22 35 63398 8966 0.141 4.11 7.30 0.681 0 0 0
Photographic and document 1954 29 36 6965 2032 0.292 0.39 3.63 0.509 2 2 0
copying equipment (MLH 351) 1958 16 19 6710 1886 0.281 0.73 4.11 0.499 2 2 0
1963 22 25 10814 2433 0.225 0.52 3.62 0.528 2 2 0
1968 19 25 40611 8106 0.200 1.10 3.19 0.567 2 2 0
1973 25 35 50657 8313 0.164 1.59 6.27 0.681 2 2 0
Watches and parts of watches 1954 8 13 6157 1364 0.222 1.20 3.49 0.509 2 0 0
1958 10 12 5312 687 0.129 1.08 3.41 0.499 2 0 0
1963 7 9 5944 1425 0.240 1.68 4.02 0.528 2 0 0
1968 10 15 11373 1330 0.117 1.89 4.32 0.567 2 0 0
Spectacles and parts of 1954 82 129 13735 2438 0.178 0.07 0.89 0.509 0 1 1
spectacles 1958 78 117 14775 3497 0.237 0.08 0.96 0.499 0 1 1
1963 76 123 17241 3081 0.179 0.09 1.13 0.528 0 1 1
1968 68 97 19438 4225 0.217 0.15 1.58 0.567 0 1 1
485
Table B6 (continued)
486
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Optical instruments 1954 21 30 9597 2589 0.270 0.39 1.42 0.509 0 2 0
1958 24 36 7776 1994 0.256 0.28 1.83 0.499 0 2 0
1963 27 43 14877 3646 0.245 0.33 1.49 0.528 0 2 0
Scientific instruments, other than 1954 169 237 70464 17754 0.252 0.31 1.58 0.509 0 2 0
optical instruments 1958 175 246 88353 22821 0.258 0.35 1.59 0.499 0 2 0
1963 171 253 125441 32006 0.255 0.51 2.09 0.528 0 2 0
Scientific and industrial 1963 247 417 200021 52256 0.261 0.34 1.42 0.528 0 2 0
instruments and systems 1968 280 464 285892 79626 0.279 0.49 1.96 0.567 0 2 0
(MLH 354)
1973 332 461 267226 67153 0.251 0.65 3.29 0.681 0 2 0
Electrical machinery (MLH 361) 1954 232 353 362912 70609 0.195 0.93 1.72 0.509 0 2 1
1958 228 373 449783 99582 0.221 1.09 1.92 0.499 0 2 1
1963 234 393 477502 75169 0.157 1.11 2.03 0.528 0 2 1
1968 220 401 485879 116319 0.239 1.27 3.13 0.567 0 2 1
1973 255 388 476458 93072 0.195 1.34 3.87 0.681 0 2 1
Insulated wires and cables 1954 50 83 163215 17718 0.109 1.95 3.45 0.509 0 1 1
(MLH 362) 1958 42 70 167783 20815 0.124 2.79 4.30 0.499 0 1 1
1963 47 91 226293 38766 0.171 2.72 4.69 0.528 0 1 1
1968 32 78 313794 46708 0.149 3.75 5.47 0.567 0 1 1
1973 34 68 328690 41434 0.126 5.12 7.63 0.681 0 1 1
Appendix B
Telegraph and telephone 1954 24 77 94159 19913 0.211 1.45 1.66 0.509 0 2 0
apparatus and equipment 1958 26 86 109117 17372 0.159 1.85 2.09 0.499 0 2 0
(MLH 363) 1963 36 86 153897 33656 0.219 2.85 2.78 0.528 0 2 0
Radio and other electronic 1954 236 407 227325 40139 0.177 0.54 1.69 0.509 0 2 0
Data Sets
equipment, except valves and 1958 250 432 297348 45764 0.154 0.62 1.80 0.499 0 2 0
semiconductors (1958 S.I.C.) 1963 271 546 492174 128292 0.261 0.65 1.75 0.528 0 2 0
Radio and electronic 1963 119 179 73362 18597 0.253 0.39 1.45 0.528 0 1 0
components, except valves and 1968 167 256 179345 52172 0.291 0.53 1.85 0.567 0 1 0
semiconductors (1968 S.I.C.)
Broadcast receiving and sound 1963 50 83 131206 22732 0.173 0.48 1.07 0.528 1 1 0
reproducing equipment, except 1968 33 66 109384 19626 0.179 0.73 1.76 0.567 1 1 0
records and tapes
Electronic computers (MLH 366) 1963 11 27 39154 12371 0.316 1.80 4.78 0.528 0 2 0
1968 20 36 92590 21326 0.230 2.24 4.33 0.567 0 2 0
1973 35 46 208460 63662 0.305 2.62 4.72 0.681 0 2 0
Radio, radar, and electronic 1963 77 168 171326 46895 0.274 1.25 2.87 0.528 0 2 0
capital goods (MLH 367) 1968 94 195 212221 49073 0.231 1.34 3.42 0.567 0 2 0
1973 115 201 289685 70163 0.242 1.61 3.75 0.681 0 2 0
Electric appliances primarily for 1954 92 140 95030 22944 0.241 0.68 2.24 0.509 2 1 1
domestic use (MLH 368) 1958 85 115 121328 23091 0.190 1.12 2.60 0.499 2 1 1
1963 102 142 211018 49486 0.235 1.16 2.32 0.528 2 1 1
1968 100 153 240736 56430 0.234 1.48 3.07 0.567 2 1 1
1973 92 119 276881 57904 0.209 2.25 4.16 0.681 2 1 1
Electrical equipment for motor 1954 33 50 63222 14327 0.227 1.11 1.96 0.509 0 1 0
vehicles, cycles, and aircraft, 1958 29 49 70874 11050 0.156 1.34 2.10 0.499 0 1 0
excluding accumulators 1963 33 63 74217 17484 0.236 1.17 2.26 0.528 0 1 0
1968 31 74 128479 35546 0.277 1.44 2.31 0.567 0 1 0
1973 31 — 138784 22775 0.164 — 2.87 0.681 0 1 0
487
Table B6 (continued)
488
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Secondary batteries 1954 14 23 24832 4827 0.194 0.50 1.46 0.509 1 0 1
(accumulators) 1958 16 21 23954 4507 0.188 0.67 1.68 0.499 1 0 1
1963 13 17 27780 5916 0.213 0.94 1.86 0.528 1 0 1
1968 18 26 40769 9680 0.237 1.28 3.03 0.567 1 0 1
Electric light fittings, wiring 1954 153 187 39233 8962 0.228 0.20 1.48 0.509 0 0 1
accessories, and other electrical 1958 168 195 59906 12937 0.216 0.23 1.38 0.499 0 0 1
goods
1963 160 192 73950 18626 0.252 0.30 1.79 0.528 0 0 1
1968 129 163 80961 20199 0.249 0.42 2.29 0.567 0 0 1
Shipbuilding and marine 1954 125 202 360296 35662 0.099 1.17 1.40 0.509 0 0 0
engineering 1958 148 216 384318 37087 0.097 1.23 1.56 0.499 0 0 0
1963 102 161 285071 32624 0.114 1.71 2.42 0.528 0 0 0
1968 81 146 223519 26045 0.117 1.56 2.26 0.567 0 0 0
Boatbuilding 1954 59 68 8992 1398 0.155 0.04 0.45 0.509 0 0 0
1958 54 62 6122 681 0.111 0.03 0.43 0.499 0 0 0
1963 46 55 7248 1079 0.149 0.02 0.33 0.528 0 0 0
1968 51 56 11329 2130 0.188 0.06 0.92 0.567 0 0 0
Cars 1954 11 25 401716 55719 0.139 27.41 9.73 0.509 1 1 0
1958 12 29 662249 75580 0.114 29.64 8.93 0.499 1 1 0
1963 15 52 1042379 170884 0.164 23.04 7.98 0.528 1 1 0
Appendix B
1968 11 43 1107976 141402 0.128 37.04 9.47 0.567 1 1 0
Motor bodies 1954 106 145 132935 15479 0.116 1.72 4.27 0.509 0 0 0
Data Sets
1958 100 131 194393 17912 0.092 2.43 4.92 0.499 0 0 0
1963 82 117 180054 22467 0.125 3.12 5.95 0.528 0 0 0
1968 87 124 150358 20159 0.134 2.08 5.09 0.567 0 0 0
Internal combustion engines for 1954 6 8 36161 5869 0.162 7.57 5.01 0.509 0 1 0
vehicles 1958 7 8 41374 3685 0.089 6.91 3.75 0.499 0 1 0
1963 6 10 59125 9306 0.157 5.47 3.32 0.528 0 1 0
Parts and accessories of motor 1954 169 219 158319 29916 0.189 2.06 7.01 0.509 0 0 0
vehicles, other than engines and 1958 142 187 182819 32714 0.179 2.18 6.25 0.499 0 0 0
motor bodies
1963 142 239 317941 62425 0.196 1.87 4.55 0.528 0 0 0
Hand tools and implements 1954 150 175 27296 5265 0.193 0.26 2.38 0.509 0 0 1
(MLH 391) 1958 121 157 30019 6930 0.231 0.29 2.46 0.499 0 0 1
1963 105 138 33132 8260 0.249 0.35 2.93 0.528 0 0 1
1968 94 122 33539 7879 0.235 0.42 3.26 0.567 0 0 1
1973 86 103 38787 10186 0.263 0.57 4.15 0.681 0 0 1
Razors and blades 1954 4 8 11773 7277 0.618 1.59 3.85 0.509 2 1 0
1958 5 7 11825 6818 0.577 2.33 5.14 0.499 2 1 0
1963 4 5 16437 8174 0.497 6.66 6.13 0.528 2 1 0
1968 4 5 20533 12372 0.603 7.66 10.17 0.567 2 1 0
Wire and wire manufactures 1954 169 231 133773 15781 0.118 0.53 3.70 0.509 0 0 1
(MLH 394) 1958 155 225 165482 18115 0.109 0.60 3.80 0.499 0 0 1
1963 141 246 184890 24667 0.133 0.72 4.51 0.528 0 0 1
1968 125 213 186584 24964 0.134 0.98 5.77 0.567 0 0 1
1973 120 193 236630 42727 0.181 1.25 6.49 0.681 0 0 1
489
Table B6 (continued)
490
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Cans and metal boxes (MLH 395) 1954 69 105 70094 8992 0.128 0.67 2.48 0.509 0 0 0
1958 63 96 82301 12013 0.146 0.89 3.12 0.499 0 0 0
1963 44 78 101514 12318 0.121 1.58 4.16 0.528 0 0 0
1968 44 75 116529 16418 0.141 2.36 6.22 0.567 0 0 0
1973 44 70 138612 15881 0.115 3.05 7.39 0.681 0 0 0
Jewellery and plate (1958 S.I.C.) 1954 199 225 25515 5400 0.212 0.08 1.15 0.509 0 0 0
1958 153 174 23848 4969 0.208 0.11 1.31 0.499 0 0 0
1963 139 166 34007 6943 0.204 0.13 1.52 0.528 0 0 0
Jewellery and plate (1968 S.I.C.) 1963 101 118 24399 4602 0.189 0.14 1.94 0.528 0 0 0
1968 67 75 29107 4800 0.165 0.38 4.45 0.567 0 0 0
Metal furniture 1954 123 142 33109 6239 0.188 0.21 1.84 0.509 0 0 0
1958 100 116 31832 5703 0.179 0.29 2.54 0.499 0 0 0
1963 118 146 50613 12094 0.239 0.27 2.01 0.528 0 0 0
1968 107 138 58593 11913 0.203 0.35 2.47 0.567 0 0 0
1973 109 133 77913 16159 0.207 0.43 2.84 0.681 0 0 0
Drop forgings of steel and iron, 1963 87 106 88475 14464 0.163 1.38 5.37 0.528 0 0 1
and steel stampings and 1968 62 91 89201 14181 0.159 1.51 5.38 0.567 0 0 1
pressings
1973 65 103 105412 17705 0.168 1.87 7.17 0.681 0 0 1
Domestic hollow-ware 1954 101 117 29105 5797 0.199 0.21 1.50 0.509 0 0 1
Appendix B
1958 84 93 19498 2363 0.121 0.25 2.03 0.499 0 0 1
1963 72 84 22866 3964 0.173 0.25 1.88 0.528 0 0 1
1968 50 61 19697 4876 0.248 0.43 3.57 0.567 0 0 1
Industrial hollow-ware 1954 81 93 27689 4795 0.173 0.32 2.94 0.509 0 0 1
Data Sets
1958 94 118 44388 6654 0.150 0.39 2.77 0.499 0 0 1
1963 60 78 32322 4926 0.152 0.66 5.31 0.528 0 0 1
1968 45 65 31439 5845 0.186 0.75 5.80 0.567 0 0 1
Metal windows and door frames 1954 34 46 26320 3254 0.124 0.46 1.64 0.509 0 0 1
1958 35 44 32013 5152 0.161 0.48 1.43 0.499 0 0 1
1963 39 56 33880 6678 0.197 0.41 1.56 0.528 0 0 1
1968 39 60 35126 8096 0.230 0.36 1.66 0.567 0 0 1
Safes, locks, latches, and keys 1963 39 45 15359 5154 0.336 0.35 1.87 0.528 0 0 1
1968 32 42 19913 5848 0.294 0.66 2.84 0.567 0 0 1
Engineers’ and mechanicians’ 1963 56 82 32464 7387 0.228 0.73 4.61 0.528 0 0 1
goods 1968 46 61 43906 9840 0.224 1.08 5.32 0.567 0 0 1
Anchors and chains 1954 37 50 9541 1729 0.181 0.24 2.69 0.509 0 0 0
1958 30 41 8906 1244 0.140 0.40 4.40 0.499 0 0 0
1963 27 33 10021 1811 0.181 0.68 7.07 0.528 0 0 0
1968 19 25 9103 1746 0.192 0.71 7.63 0.567 0 0 0
Spun cotton yarns 1954 136 287 190888 20326 0.106 0.65 2.35 0.752 0 0 1
1958 102 231 104674 9038 0.086 0.85 3.27 0.720 0 0 1
1963 51 111 64778 6175 0.095 1.80 6.03 0.790 0 0 1
1968 40 77 52826 4553 0.086 2.12 7.92 0.810 0 0 1
Cotton waste yarns 1954 48 68 16087 2799 0.174 0.23 2.87 0.752 0 0 1
1958 45 61 11765 2075 0.176 0.25 2.94 0.720 0 0 1
1963 31 44 10170 1275 0.125 0.33 3.02 0.790 0 0 1
1968 19 25 4273 480 0.112 0.40 4.60 0.810 0 0 1
491
Table B6 (continued)
492
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Spun man-made fibres and 1954 32 40 21005 3105 0.148 0.63 2.54 0.752 0 0 1
mixture yarns 1958 32 42 20354 1263 0.062 0.59 2.32 0.720 0 0 1
1963 24 37 20716 2470 0.119 1.02 4.11 0.790 0 0 1
1968 18 36 33794 4877 0.144 1.59 5.15 0.810 0 0 1
Creped, bulked, textured, or 1963 26 38 26982 3552 0.132 2.11 11.37 0.790 0 0 0
stretch continuous filament yarns 1968 27 53 66673 14572 0.219 4.00 13.94 0.810 0 0 0
Finished thread 1954 31 45 30305 4211 0.139 0.65 1.97 0.752 0 0 0
1958 24 38 24631 3851 0.156 1.19 4.07 0.720 0 0 0
1963 24 37 23646 3847 0.163 2.18 7.97 0.790 0 0 0
1968 18 28 29301 9796 0.334 2.77 8.86 0.810 0 0 0
Loom state cloth of cotton 1954 402 556 174219 13237 0.076 0.28 2.19 0.752 0 0 0
1958 252 362 97130 6347 0.065 0.44 3.30 0.720 0 0 0
1963 119 164 54511 4164 0.076 0.61 4.59 0.790 0 0 0
1968 73 99 34264 2872 0.084 0.70 5.64 0.810 0 0 0
Loom state cloth of man-made 1954 162 209 77785 7233 0.093 0.52 3.20 0.752 0 0 1
fibres and mixtures 1958 110 149 56182 4269 0.076 0.62 4.12 0.720 0 0 1
1963 69 102 62504 4982 0.080 1.24 6.86 0.790 0 0 1
1968 52 78 52929 5997 0.113 2.33 14.58 0.810 0 0 1
Wool combers (on own account) 1954 10 12 31294 1867 0.060 0.45 1.82 0.320 0 0 0
Appendix B
1958 14 16 28775 2226 0.077 0.64 3.02 0.314 0 0 0
1963 18 35 47279 4627 0.098 0.70 4.09 0.310 0 0 0
Wool combers (on commission) 1954 39 59 13443 4497 0.335 0.49 3.00 0.320 0 0 1
Data Sets
1958 34 55 12858 3567 0.277 0.80 4.63 0.314 0 0 1
1963 26 47 14971 4647 0.310 0.93 5.15 0.310 0 0 1
Spun woollen yarns 1954 91 103 44201 4728 0.107 0.34 3.78 0.320 0 0 0
1958 85 103 38705 5130 0.133 0.34 3.65 0.314 0 0 0
1963 83 108 61631 11066 0.180 0.46 3.46 0.310 0 0 0
1968 80 108 60494 11362 0.188 0.81 6.05 0.346 0 0 0
Woven worsted fabric 1954 222 301 123535 16119 0.130 0.26 1.96 0.320 0 0 0
1958 182 252 96368 10333 0.107 0.28 2.13 0.314 0 0 0
1963 130 229 95805 13226 0.138 0.34 2.23 0.310 0 0 0
1968 83 171 73266 12630 0.172 0.49 3.49 0.346 0 0 0
Woven woollen fabric, other 1954 253 324 116771 18863 0.162 0.34 2.08 0.320 0 0 0
than blankets, traveling rugs, etc. 1958 217 308 95201 11959 0.126 0.31 1.97 0.314 0 0 0
1963 162 237 78752 15308 0.194 0.36 2.27 0.310 0 0 0
1968 135 200 73924 14532 0.197 0.52 3.23 0.346 0 0 0
Mechanical cloth, woven felt, 1954 20 33 15088 1527 0.101 0.54 3.78 0.320 0 0 1
and pressed felt 1958 16 27 11508 2611 0.227 0.59 3.90 0.314 0 0 1
1963 13 20 8972 2232 0.249 0.74 4.88 0.310 0 0 1
1968 13 22 11636 3318 0.285 0.97 5.95 0.346 0 0 1
Jute (MLH 415) 1954 47 72 43789 6016 0.137 0.86 3.20 0.320 0 0 1
1958 48 74 38026 3252 0.086 0.91 4.14 0.314 0 0 1
1963 32 64 38058 5190 0.136 1.02 4.03 0.310 0 0 1
1968 28 57 39104 6035 0.154 1.24 5.03 0.346 0 0 1
1973 21 40 39220 6997 0.178 1.91 7.67 0.449 0 0 1
493
Table B6 (continued)
494
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Rope, twine, and net (MLH 416) 1954 82 127 32957 4595 0.139 0.21 1.94 0.320 0 0 1
1958 64 110 27823 4633 0.167 0.28 2.55 0.314 0 0 1
1963 47 99 31223 5564 0.178 0.34 3.14 0.310 0 0 1
1968 34 54 23690 4720 0.199 0.70 4.73 0.346 0 0 1
1973 23 40 22809 4383 0.192 1.09 7.12 0.449 0 0 1
Hosiery 1954 240 321 105983 22583 0.213 0.94 5.85 0.320 1 0 0
1958 188 248 79577 13304 0.167 1.14 6.98 0.314 1 0 0
1963 157 202 65001 13436 0.207 1.04 6.92 0.310 1 0 0
1968 92 148 78294 17925 0.229 1.06 5.36 0.346 1 0 0
Underwear, shirts, and 1954 100 144 38471 6014 0.156 0.23 1.59 0.320 0 0 0
nightwear, knitted, netted, or 1958 90 139 38043 6165 0.162 0.33 2.20 0.314 0 0 0
crocheted 1963 58 114 38649 6750 0.175 0.49 2.81 0.310 0 0 0
1968 52 101 50444 12221 0.242 0.60 3.00 0.346 0 0 0
Other garments, etc., knitted, 1954 326 408 62220 11085 0.178 0.15 1.65 0.320 0 0 0
netted, or crocheted 1958 293 379 70102 11562 0.165 0.28 2.54 0.314 0 0 0
1963 257 394 114945 21303 0.185 0.49 3.31 0.310 0 0 0
1968 245 410 138644 27366 0.197 0.64 4.17 0.346 0 0 0
Woven carpets, carpeting, and 1954 52 84 76253 14337 0.188 0.99 2.94 0.320 0 0 1
carpet floor rugs 1958 54 91 78295 14358 0.183 0.94 2.98 0.314 0 0 1
Appendix B
1963 42 68 79534 16127 0.203 1.24 3.22 0.310 0 0 1
1968 32 68 80275 15588 0.194 1.25 3.67 0.346 0 0 1
Tufted carpets, carpeting, and 1958 5 7 4748 618 0.130 0.90 5.98 0.314 1 0 0
Data Sets
carpet floor rugs 1963 20 24 21159 3938 0.186 0.85 5.20 0.310 1 0 0
1968 27 37 56536 6529 0.115 1.44 5.41 0.346 1 0 0
Made-up household textiles and 1954 215 244 34227 3645 0.106 0.04 0.49 0.320 0 0 0
handkerchiefs (MLH 422.1) 1958 204 254 35580 4518 0.127 0.05 0.65 0.314 0 0 0
1963 190 232 41617 5673 0.136 0.07 0.89 0.310 0 0 0
1968 147 186 46617 6591 0.141 0.14 1.60 0.346 0 0 0
1973 120 194 69003 14772 0.214 0.23 2.34 0.449 0 0 0
Canvas sacks and bags 1954 80 115 30866 1582 0.051 0.05 0.90 0.320 0 0 1
1958 68 95 20322 2217 0.109 0.06 1.15 0.314 0 0 1
1963 47 71 18744 1247 0.067 0.06 1.11 0.310 0 0 1
1968 31 39 7558 777 0.103 0.08 1.64 0.346 0 0 1
Textile finishing: yarn 1954 92 111 13542 2238 0.165 0.27 2.93 0.320 0 0 1
1958 70 87 10760 1835 0.171 0.32 3.67 0.314 0 0 1
1963 52 68 10634 2594 0.244 0.50 4.86 0.310 0 0 1
1968 41 56 10861 2770 0.255 0.86 8.49 0.346 0 0 1
Textile finishing: woven fabrics 1954 179 241 64563 13743 0.213 0.47 2.61 0.320 0 0 1
of cotton and man-made fibres 1958 151 220 58328 8681 0.149 0.65 3.58 0.314 0 0 1
1963 96 167 48969 10003 0.204 0.90 5.04 0.310 0 0 1
1968 76 123 42515 10183 0.240 1.28 7.39 0.346 0 0 1
Textile finishing: woollen and 1954 60 71 9624 2534 0.263 0.36 4.44 0.320 0 0 1
worsted fabrics 1958 45 52 6744 1836 0.272 0.38 5.50 0.314 0 0 1
1963 37 45 6617 1658 0.251 0.30 3.64 0.310 0 0 1
1968 23 29 4783 1282 0.268 0.49 5.84 0.346 0 0 1
495
Table B6 (continued)
496
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Textile finishing: knitted fabrics 1954 64 75 10559 2907 0.275 0.36 2.91 0.320 0 0 1
and other knitted goods 1958 51 63 10904 2854 0.262 0.52 3.74 0.314 0 0 1
1963 44 57 15408 4580 0.297 0.87 4.94 0.310 0 0 1
1968 50 70 23572 7396 0.314 1.08 5.93 0.346 0 0 1
Asbestos manufactures 1954 29 40 42462 9203 0.217 1.18 2.95 0.320 0 0 1
(MLH 429.1) 1958 30 41 45229 11116 0.246 1.59 3.60 0.314 0 0 1
1963 23 40 55278 13314 0.241 1.80 3.60 0.310 0 0 1
1968 20 39 66278 15944 0.241 2.24 4.29 0.346 0 0 1
1973 16 33 65727 14066 0.214 3.01 5.55 0.449 0 0 1
Needle felt and needleloom 1954 13 17 5514 1168 0.212 0.14 2.20 0.320 0 0 0
carpeting 1958 13 21 4734 713 0.151 0.14 2.50 0.314 0 0 0
1963 15 25 7124 1277 0.179 0.19 2.94 0.310 0 0 0
1968 14 24 8298 2177 0.262 0.40 5.02 0.346 0 0 0
Leather tanning 1954 72 85 38382 2754 0.072 0.15 1.50 0.330 0 0 0
1958 49 56 21405 1666 0.078 0.20 2.05 0.323 0 0 0
1963 29 32 17045 1938 0.114 0.31 2.63 0.318 0 0 0
1968 22 25 15512 1783 0.115 0.50 4.13 0.284 0 0 0
Leather tanning and dressing, or 1954 170 194 65312 6249 0.096 0.22 2.33 0.330 0 0 0
dressing only 1958 150 172 56692 5827 0.103 0.23 2.43 0.323 0 0 0
Appendix B
1963 130 156 59245 8234 0.139 0.27 2.82 0.318 0 0 0
1968 120 140 67775 11078 0.163 0.36 3.63 0.284 0 0 0
Fellmongery 1954 25 27 11148 1103 0.099 0.13 2.48 0.330 0 0 0
Data Sets
1958 22 24 8095 549 0.068 0.16 2.90 0.323 0 0 0
1963 29 37 17194 1588 0.092 0.14 2.68 0.318 0 0 0
1968 20 24 11301 1117 0.099 0.22 3.89 0.284 0 0 0
Travel goods of leather or leather 1954 50 56 8974 1632 0.182 0.06 0.61 0.330 0 0 0
substitutes 1958 33 38 6808 1247 0.183 0.08 0.81 0.323 0 0 0
1963 31 36 7735 1557 0.201 0.11 0.99 0.318 0 0 0
1968 38 45 10300 1823 0.177 0.11 0.99 0.284 0 0 0
Leather goods other than travel 1954 157 176 13442 2024 0.151 0.02 0.36 0.330 0 0 0
goods 1958 125 134 11890 2042 0.172 0.03 0.49 0.323 0 0 0
1963 120 131 16553 2965 0.179 0.05 0.75 0.318 0 0 0
1968 110 130 16823 3272 0.194 0.06 0.92 0.284 0 0 0
Fur (MLH 433) 1954 72 84 9069 1554 0.171 0.08 1.39 0.330 0 0 0
1958 52 55 8614 1158 0.134 0.11 1.62 0.323 0 0 0
1963 53 65 15239 3253 0.213 0.11 1.49 0.318 0 0 0
1968 46 55 13112 2123 0.162 0.14 2.14 0.284 0 0 0
1973 41 57 11104 2121 0.191 0.15 2.69 0.267 0 0 0
Weatherproof outerwear (MLH 1954 — 304 45817 5898 0.129 0.07 0.69 0.337 0 0 0
441) 1958 180 242 35686 4591 0.129 0.08 0.83 0.326 0 0 0
1963 157 232 44065 7353 0.167 0.09 0.82 0.319 0 0 0
1968 118 186 35433 7102 0.200 0.10 1.06 0.333 0 0 0
1973 120 178 34884 6999 0.201 0.12 1.29 0.369 0 0 0
497
Table B6 (continued)
498
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Men’s and boys’ tailored 1954 544 760 168227 21008 0.125 0.05 0.33 0.337 0 0 0
outerwear (MLH 442) 1958 459 652 153941 17986 0.117 0.07 0.39 0.326 0 0 0
1963 352 592 163246 23884 0.146 0.09 0.50 0.319 0 0 0
1968 302 551 161643 27406 0.170 0.12 0.72 0.333 0 0 0
1973 315 513 180010 34679 0.193 0.16 0.93 0.369 0 0 0
Women’s and girls’ tailored 1954 521 635 86800 11507 0.133 0.05 0.60 0.337 0 0 0
outerwear (MLH 443) 1958 406 520 74984 9193 0.123 0.06 0.66 0.326 0 0 0
1963 369 465 83662 12967 0.155 0.07 0.75 0.319 0 0 0
1968 279 382 80623 13760 0.171 0.10 1.02 0.333 0 0 0
1973 355 416 87260 18909 0.217 0.10 1.15 0.369 0 0 0
Heavy overalls and jeans 1954 95 127 17596 1751 0.100 0.05 0.59 0.337 0 0 0
1958 76 99 16160 1871 0.116 0.07 0.65 0.326 0 0 0
1963 81 114 20516 3469 0.169 0.06 0.66 0.319 0 0 0
1968 73 112 21566 3769 0.175 0.08 0.76 0.333 0 0 0
Shirts, underwear, etc. (on own 1954 213 273 44879 5099 0.114 0.05 0.44 0.337 1 0 0
account) 1958 181 246 45719 5243 0.115 0.06 0.51 0.326 1 0 0
1963 149 227 52713 7485 0.142 0.09 0.72 0.319 1 0 0
1968 111 183 51955 8750 0.168 0.15 1.06 0.333 1 0 0
Women’s and girls’ light 1954 549 665 68151 8938 0.131 0.03 0.43 0.337 0 0 0
Appendix B
outerwear (on own account) 1958 492 611 69599 9010 0.129 0.03 0.46 0.326 0 0 0
1963 329 453 63530 10908 0.172 0.05 0.65 0.319 0 0 0
1968 272 403 86700 14397 0.166 0.07 0.74 0.333 0 0 0
Women’s and girls’ underwear 1954 148 190 24229 2692 0.111 0.05 0.57 0.337 0 0 0
Data Sets
and nightwear 1958 146 186 30385 2642 0.087 0.06 0.62 0.326 0 0 0
1963 122 172 33041 4452 0.135 0.10 0.86 0.319 0 0 0
1968 118 186 33513 6074 0.181 0.13 1.36 0.333 0 0 0
Infants’ wear 1954 127 149 12564 2029 0.161 0.03 0.41 0.337 0 0 0
1958 119 152 14414 2316 0.161 0.03 0.49 0.326 0 0 0
1963 94 132 13504 2195 0.163 0.06 0.78 0.319 0 0 0
1968 93 140 18789 3629 0.193 0.08 1.04 0.333 0 0 0
Corsetry 1954 69 105 21817 4559 0.209 0.20 1.40 0.337 2 0 0
1958 62 101 28208 6340 0.225 0.22 1.27 0.326 2 0 0
1963 63 123 36944 8763 0.237 0.19 1.19 0.319 2 0 0
1968 53 119 40566 10123 0.250 0.20 1.25 0.333 2 0 0
Gloves (MLH 449.2) 1954 120 188 14406 2688 0.187 0.03 0.59 0.337 0 0 0
1958 89 139 10505 1798 0.171 0.04 0.80 0.326 0 0 0
1963 76 122 12731 2387 0.187 0.05 0.84 0.319 0 0 0
1968 55 92 11338 2077 0.183 0.07 1.11 0.333 0 0 0
1973 55 80 12770 2627 0.206 0.09 1.45 0.369 0 0 0
Footwear (MLH 450) 1954 529 778 192337 28566 0.149 0.09 0.63 0.718 1 0 0
1958 457 674 176126 25157 0.143 0.11 0.72 0.724 1 0 0
1963 370 615 206016 37473 0.182 0.15 0.91 0.731 1 0 0
1968 299 559 207150 38775 0.187 0.20 1.21 0.779 1 0 0
1973 266 441 238701 47744 0.200 0.29 1.54 0.813 1 0 0
Refractory goods 1954 88 147 33677 5805 0.172 0.25 2.25 0.268 0 0 1
1958 73 149 37460 6518 0.174 0.42 3.95 0.257 0 0 1
1963 63 130 40633 8734 0.215 0.63 6.02 0.256 0 0 1
499
1968 44 111 52710 12125 0.230 0.71 5.67 0.289 0 0 1
1973 47 — 60518 16273 0.269 — 6.56 0.385 0 0 1
Table B6 (continued)
500
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Building bricks 1954 309 543 60779 12568 0.207 0.12 1.70 0.268 0 0 1
1958 224 525 54900 9440 0.172 0.12 1.87 0.257 0 0 1
1963 179 433 71157 16717 0.235 0.20 2.49 0.256 0 0 1
1968 109 324 64763 15739 0.243 0.45 5.25 0.289 0 0 1
Sanitary ware of fireclay, etc. 1954 82 97 17909 3339 0.186 0.14 1.15 0.268 0 0 1
1958 70 85 13784 2066 0.150 0.19 1.75 0.257 0 0 1
1963 60 81 17383 3606 0.207 0.26 2.34 0.256 0 0 1
1968 35 72 17796 3823 0.215 0.42 3.88 0.289 0 0 1
Tiles, pipes, and other products 1954 62 84 9536 1714 0.180 0.17 1.91 0.268 0 0 1
of fireclay, etc. 1958 52 96 9312 1384 0.149 0.13 1.96 0.257 0 0 1
1963 32 51 9728 2302 0.237 0.24 2.53 0.256 0 0 1
1968 28 37 11420 2728 0.239 0.45 3.79 0.289 0 0 1
Tiles, other than of precast 1954 36 44 12799 2808 0.219 0.54 2.36 0.574 0 0 0
concrete and brick earth 1958 25 32 11454 2234 0.195 0.79 3.17 0.524 0 0 0
1963 17 28 16758 3863 0.231 1.52 5.04 0.569 0 0 0
1968 10 22 19210 3499 0.182 2.78 8.09 0.721 0 0 0
Sanitary earthenware 1954 16 21 7719 2117 0.274 0.28 1.33 0.574 0 0 1
1958 14 17 6663 1502 0.225 0.59 2.34 0.524 0 0 1
1963 14 18 11562 2818 0.244 0.80 2.46 0.569 0 0 1
Appendix B
1968 7 14 12762 3348 0.262 0.74 2.20 0.721 0 0 1
China, earthenware, etc., 1954 117 143 30243 6367 0.211 0.29 1.11 0.574 0 0 1
Data Sets
domestic and ornamental 1958 116 152 32197 6611 0.205 0.26 1.05 0.524 0 0 1
1963 94 132 35215 7119 0.202 0.26 0.95 0.569 0 0 1
1968 76 121 40676 9435 0.232 0.32 1.15 0.721 0 0 1
Electrical ware of porcelain, 1954 14 20 6562 1456 0.222 0.58 1.95 0.574 0 0 1
earthenware, or stoneware 1958 12 18 8122 2027 0.250 0.83 2.53 0.524 0 0 1
1963 13 22 10133 2653 0.262 1.15 3.58 0.569 0 0 1
1968 10 20 10413 1995 0.192 1.23 3.85 0.721 0 0 1
Glass containers 1954 38 58 41369 7882 0.191 1.22 2.98 0.411 0 0 1
1958 32 51 48845 9591 0.196 1.69 3.49 0.394 0 0 1
1963 23 41 56727 10781 0.190 2.27 3.91 0.422 0 0 1
1968 19 37 62587 13997 0.224 3.02 5.15 0.524 0 0 1
Domestic and fancy glassware 1954 19 26 7062 1858 0.263 0.11 0.47 0.411 1 0 0
1958 16 20 7544 1563 0.207 0.29 1.00 0.394 1 0 0
1963 15 18 9446 3088 0.327 0.41 1.32 0.422 1 0 0
1968 16 24 18584 5410 0.291 0.67 1.67 0.524 1 0 0
Glass fibre and miscellaneous 1954 15 22 6283 1995 0.318 0.55 4.49 0.411 0 1 0
glass products 1958 22 34 10631 3104 0.292 0.43 3.40 0.394 0 1 0
1963 29 40 14873 4490 0.302 0.66 5.10 0.422 0 1 0
Glass fibre and manufactures 1963 9 12 9683 3334 0.344 2.45 9.56 0.422 0 2 0
thereof 1968 7 10 12907 5270 0.408 3.22 9.15 0.524 0 2 0
Cement (MLH 464) 1954 13 47 68489 19738 0.288 2.52 9.41 0.268 0 0 0
1958 13 50 70011 17720 0.253 2.71 10.64 0.257 0 0 0
1963 14 54 83280 26467 0.318 3.79 14.55 0.256 0 0 0
1968 9 53 89608 25912 0.289 6.26 24.08 0.289 0 0 0
501
1973 8 53 127577 39435 0.309 7.45 27.07 0.385 0 0 0
Table B6 (continued)
502
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Lime and whiting 1954 37 45 9671 2038 0.211 0.60 6.71 0.268 0 0 1
1958 33 49 9989 2392 0.239 0.63 8.74 0.257 0 0 1
1963 26 35 9999 2776 0.278 0.71 8.92 0.256 0 0 1
1968 11 17 11626 2421 0.208 2.53 20.61 0.289 0 0 1
Pre-cast concrete goods 1954 258 362 48767 8572 0.176 0.14 2.01 0.268 0 0 1
1958 206 312 49582 8490 0.171 0.19 2.63 0.257 0 0 1
1963 201 337 65103 13518 0.208 0.23 3.06 0.256 0 0 1
1968 229 384 107266 25685 0.239 0.41 5.04 0.289 0 0 1
Asbestos cement goods 1954 8 20 21662 4755 0.220 1.18 3.11 0.268 0 0 1
1958 7 21 20931 4575 0.219 1.25 3.61 0.257 0 0 1
1963 8 23 26577 6613 0.249 1.47 3.94 0.256 0 0 1
1968 8 22 27503 7212 0.262 2.06 5.53 0.289 0 0 1
Builders’ woodwork and 1954 180 202 39392 4570 0.116 0.09 0.89 0.456 0 0 1
prefabricated timber buildings 1958 147 173 40900 5089 0.124 0.12 1.11 0.435 0 0 1
1963 173 208 72485 14250 0.197 0.16 1.38 0.394 0 0 1
1968 228 282 103423 19090 0.185 0.18 1.70 0.323 0 0 1
Wood chipboard 1963 9 16 6603 1387 0.210 0.78 6.17 0.394 0 0 0
1968 11 18 11087 2166 0.195 1.04 6.39 0.323 0 0 0
Plywood and textured boards 1963 17 17 7218 569 0.079 0.41 3.17 0.394 0 0 0
Appendix B
1968 19 24 17348 2560 0.148 0.43 3.41 0.323 0 0 0
Domestic furniture, not 1954 340 424 90625 12637 0.139 0.11 0.98 0.456 0 0 0
Data Sets
upholstered, mainly of wood 1958 265 389 81017 11896 0.147 0.12 1.06 0.435 0 0 0
1963 238 380 89233 15734 0.176 0.13 1.16 0.394 0 0 0
1968 237 298 111984 25856 0.231 0.24 1.98 0.323 0 0 0
Upholstered furniture 1954 207 249 41073 6001 0.146 0.03 0.43 0.456 0 0 0
1958 155 190 41476 5678 0.137 0.05 0.54 0.435 0 0 0
1963 159 195 55426 9830 0.177 0.07 0.70 0.394 0 0 0
1968 162 200 56159 10789 0.192 0.10 1.13 0.323 0 0 0
Office, school, and other 1954 174 199 28139 4339 0.154 0.08 0.91 0.456 0 0 0
furniture 1958 179 216 36409 5271 0.145 0.10 1.04 0.435 0 0 0
1963 137 162 33916 5949 0.175 0.17 1.73 0.394 0 0 0
1968 112 132 36408 6781 0.186 0.20 1.93 0.323 0 0 0
Bedding 1954 108 145 28579 4078 0.143 0.15 1.94 0.456 1 0 0
1958 72 103 26141 3686 0.141 0.19 2.13 0.435 1 0 0
1963 55 53 30032 5471 0.182 0.35 2.06 0.394 1 0 0
1968 53 90 44569 8526 0.191 0.20 1.63 0.323 1 0 0
Wooden containers and baskets 1954 271 332 39675 4607 0.116 0.05 0.87 0.456 0 0 0
(MLH 475) 1958 192 229 29239 3828 0.131 0.08 1.34 0.435 0 0 0
1963 165 216 30027 4554 0.152 0.09 1.55 0.394 0 0 0
1968 147 197 31349 5388 0.172 0.11 1.92 0.323 0 0 0
1973 160 228 49112 10133 0.206 0.12 2.27 0.340 0 0 0
Cork manufactures 1954 19 25 3022 575 0.190 0.10 1.71 0.456 0 0 0
1958 13 18 3348 576 0.172 0.25 3.39 0.435 0 0 0
1963 8 9 3855 1013 0.263 0.59 3.87 0.394 0 0 0
1968 7 10 4116 812 0.197 0.47 3.98 0.323 0 0 0
503
Table B6 (continued)
504
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Paper and board (MLH 481) 1954 186 253 311039 71029 0.228 1.83 6.37 0.367 0 0 1
1958 167 262 332236 57771 0.174 2.74 8.81 0.361 0 0 1
1963 144 267 381309 74910 0.196 3.18 9.66 0.377 0 0 1
1968 111 207 385640 71760 0.186 5.05 13.98 0.403 0 0 1
1973 92 171 415335 74153 0.179 6.31 18.04 0.481 0 0 1
Rigid boxes 1954 211 252 25264 4885 0.193 0.12 1.54 0.367 0 0 0
1958 177 216 21835 4041 0.185 0.13 1.65 0.361 0 0 0
1963 154 188 21606 4474 0.207 0.12 1.50 0.377 0 0 0
1968 117 147 19616 3891 0.198 0.16 2.05 0.403 0 0 0
Cartons 1954 71 92 42532 7823 0.184 0.79 3.68 0.367 0 0 0
1958 72 91 52073 7774 0.149 1.07 4.35 0.361 0 0 0
1963 75 102 66679 10795 0.162 1.25 5.15 0.377 0 0 0
1968 83 113 90021 16757 0.186 1.32 5.47 0.403 0 0 0
Fibreboard packing cases 1954 27 43 38845 8289 0.213 0.94 4.07 0.367 0 0 0
1958 26 45 46420 8211 0.177 1.03 4.28 0.361 0 0 0
1963 46 77 75703 13950 0.184 0.92 3.79 0.377 0 0 0
1968 49 106 112448 18763 0.167 1.18 4.90 0.403 0 0 0
Composite containers of board 1963 10 12 5046 905 0.179 0.64 4.68 0.377 0 0 0
and metal 1968 8 9 6120 1250 0.204 0.81 3.04 0.403 0 0 0
Appendix B
Packaging products of paper and 1963 18 22 12122 3338 0.275 0.75 5.54 0.377 0 0 0
associated materials other than 1968 22 36 28625 4477 0.156 0.99 5.98 0.403 0 0 0
bags and sacks
Notepaper, envelopes, and 1954 38 50 23205 5164 0.223 0.50 2.17 0.367 0 0 1
Data Sets
boxed stationery, plain 1958 40 58 31337 6112 0.195 0.52 1.97 0.361 0 0 1
postcards, and letter cards 1963 29 44 30927 7668 0.248 0.91 3.18 0.377 0 0 1
Magazines and periodicals 1954 104 146 107087 23731 0.222 0.43 1.77 0.838 1 0 0
1958 91 128 127585 24462 0.192 0.66 2.22 0.838 1 0 0
1963 89 149 151136 38114 0.252 0.69 2.70 0.835 1 0 0
1968 101 170 149372 37386 0.250 0.75 3.67 0.793 1 0 0
Books 1954 127 139 54363 12424 0.229 0.35 2.10 0.838 1 0 0
1958 131 151 70757 15928 0.225 0.48 2.44 0.838 1 0 0
1963 113 149 89774 27755 0.309 0.76 3.91 0.835 1 0 0
Stereotyping, electrotyping, and 1954 132 167 15780 3569 0.226 0.13 1.95 0.838 0 0 1
engraving 1958 141 175 20284 4431 0.218 0.16 2.32 0.838 0 0 1
1963 141 198 25790 5856 0.227 0.20 2.89 0.835 0 0 1
Cellular rubber products 1954 7 9 11450 1369 0.120 2.14 4.21 0.304 0 0 0
1958 9 12 11155 1056 0.095 1.58 5.72 0.309 0 0 0
1963 11 16 15573 2260 0.145 1.18 3.96 0.344 0 0 0
Rubber products other than 1963 145 223 138082 28431 0.206 0.75 3.00 0.344 0 0 0
tyres, hose and tubing, and 1968 133 213 156152 38990 0.250 0.94 3.70 0.289 0 0 0
belting
Linoleum, plastic floor covering, 1954 22 29 50166 8403 0.167 2.12 3.97 0.304 0 0 1
leather cloth, etc. (MLH 492) 1958 21 31 47266 8095 0.171 2.25 4.75 0.309 0 0 1
1963 29 44 77154 17841 0.231 2.27 5.00 0.344 0 0 1
1968 21 31 71197 17027 0.239 3.81 6.97 0.289 0 0 1
1973 23 37 91327 25859 0.283 3.72 8.51 0.308 0 0 1
505
Table B6 (continued)
506
Industry Year NFIRMS NPLANTS SS PROFIT PCM K/N K/L UNION ADTYPE RDTYPE CHANGE
Brushes and brooms (MLH 493) 1954 116 143 20781 3529 0.170 0.13 1.51 0.304 1 0 0
1958 85 106 19241 4574 0.238 0.18 1.84 0.309 1 0 0
1963 66 85 20905 4226 0.202 0.25 2.23 0.344 1 0 0
1968 50 70 20913 4159 0.199 0.37 3.00 0.289 1 0 0
1973 60 70 24061 5528 0.230 0.45 3.66 0.308 1 0 0
Toys and games 1954 107 140 35143 6665 0.190 0.22 1.30 0.304 1 0 0
1958 87 117 35248 7988 0.227 0.29 1.69 0.309 1 0 0
1963 69 108 39519 10791 0.273 0.40 2.05 0.344 1 0 0
1968 64 104 57697 17575 0.305 0.66 2.55 0.289 1 0 0
Children’s and invalids’ 1954 23 25 5818 949 0.163 0.08 0.58 0.509 1 0 0
carriages 1958 17 19 4898 802 0.164 0.12 0.80 0.499 1 0 0
1963 21 27 7484 1390 0.186 0.15 1.06 0.528 1 0 0
1968 13 19 8401 1962 0.234 0.33 1.84 0.567 1 0 0
Pens and mechanical pencils 1954 28 35 7560 2110 0.279 0.36 2.93 0.304 2 0 0
1958 18 23 9454 3147 0.333 0.46 2.36 0.309 2 0 0
1963 14 18 9709 3428 0.353 0.56 2.86 0.344 2 0 0
Lead pencils, crayons, etc. 1958 12 12 2428 623 0.257 0.24 1.94 0.309 1 0 1
1963 11 12 2512 614 0.244 0.34 2.97 0.344 1 0 1
Office machinery requisites 1954 18 26 8343 2078 0.249 0.27 1.73 0.304 0 0 0
Appendix B
1958 19 25 10283 2118 0.206 0.29 1.37 0.309 0 0 0
1963 18 26 13234 2909 0.220 0.46 2.17 0.344 0 0 0
1968 21 28 19115 5384 0.282 0.68 3.84 0.289 0 0 0
Plastics products (MLH 496) 1954 218 258 45274 8217 0.182 0.22 1.95 0.304 0 0 0
Data Sets
1958 234 295 65068 11416 0.175 0.25 2.14 0.309 0 0 0
1963 377 491 145292 33741 0.232 0.32 2.63 0.344 0 0 0
1968 472 625 251331 62572 0.249 0.57 4.29 0.289 0 0 0
1973 645 722 394152 97511 0.247 0.85 6.26 0.308 0 0 0
Musical instruments 1954 51 63 5719 953 0.167 0.08 1.27 0.304 0 0 0
1958 39 50 5549 980 0.177 0.08 1.10 0.309 0 0 0
1963 30 42 6829 1684 0.247 0.08 0.83 0.344 0 0 0
1968 30 48 8285 1858 0.224 0.12 1.68 0.289 0 0 0
1973 34 52 13468 3323 0.247 0.14 1.85 0.308 0 0 0
Notes: The classification according to MLH industries follows the 1968 Standard Industrial Classification, unless otherwise stated.
NFIRMS is the number of firms employing at least 25 persons.
NPLANTS is the number of plants employing at least 25 persons.
SS is sales revenue (in £1000) deflated by the general producer price index.
PROFIT is the net value of output minus wages and salaries (in £1000) deflated by the general producer price index.
PCM is the price-cost margin, i.e., net value of output minus wages and salaries divided by sales revenue.
K is the value of capital stock in 1980 prices (in £ million) (my estimate; see text for details).
L is employment (in 1000).
N is the number of plants with at least 25 employees.
UNION is union density (i.e., the number of unionized employees over the total number of employees) of the corresponding industry group.
ADTYPE takes the value 0 for an industry with typical advertising-sales ratio lower than 1%, 1 for an industry with typical advertising-sales
ratio between 1% and 2%, and 2 for an industry with typical advertising-sales ratio higher than 2%.
RDTYPE takes the value 0 for an industry with typical R&D-sales ratio lower than 1%, 1 for an industry with typical R&D-sales ratio between
1% and 2%, and 2 for an industry with typical R&D-sales ratio higher than 2%.
CHANGE takes the value 1 for an industry with a change in competition regime and 0 otherwise.
507
This page intentionally left blank
Glossary of British Terms
Appendix A
UK US
Aluminium Aluminum
Animal feeding stuffs Animal foodstuffs
Children’s carriages Baby carriages
Electric cookers Electric stoves
Fibreboard packing cases Fiberboard packing cases
Gelatine Gelatin
Glass fibres Glass fibers
Gramophone records Phonograph records
Jewellery Jewelry
Lifts Elevators
Made-up textiles Fully manufactured textiles
Man-made fibres Man-made fibers
Pedal cycles Bicycles
Perry Fermented pear juice
Petrol Gasoline
Self-raising flour Self-rising flour
Slide fasteners Zippers
Spectacles Eyeglasses
Trams Streetcars
Tyres Tires
Wagons Railway freight cars
510 Glossary
Whisky Whiskey
Woollen Woolen
Appendix B
UK US
Accumulators Storage batteries
Aluminium Aluminium
Carbons Carbon paper
Children’s carriages Baby carriages
Electric cookers Electric stoves
Engineers’ and mechanicians’ goods Engineers’ and mechanics’ goods
Fibreboard packing cases Fiberboard packing cases
Gelatine Gelatin
Glass fibres Glass fibers
Gramophone records Phonograph records
Gramophones Phonographs
Ice lollies Ice pops
Invalids’ carriages Wheelchairs
Jewellery Jewelry
Lifts Elevators
Made-up textiles Fully manufactured textiles
Man-made fibres Man-made fibers
Perry Fermented pear juice
Potato crisps Potato chips
Pyjamas Pajamas
Sanitary towels Sanitary napkins
Spectacles Eyeglasses
Tyres Tires
Whisky Whiskey
Woollen Woolen
References
Publications used only as data sources are detailed in the text and are not included here.
Acs, Z. J., and D. B. Audretsch (1988). Innovation in large and small firms: An empir-
ical analysis. American Economic Review, 78: 678–690.
Aghion, P., M. Dewatripont, and P. Rey (1999). Competition, financial discipline and
growth. Review of Economic Studies, 66: 825–852.
Aghion, P., and P. Howitt (1997). A Schumpeterian perspective on growth and com-
petition. In D. M. Kreps and K. F. Wallis, eds., Advances in Economics and Econometrics:
Theory and Applications. Seventh World Congress. Vol. 2. Cambridge: Cambridge Uni-
versity Press.
Aghion, P., and P. Howitt (1998). Endogenous Growth Theory. Cambridge, MA: M.I.T.
Press.
Aghion, P., and M. Schankerman (2000). A Model of Market Enhancing Infrastructure.
CEPR Discussion Paper no. 2462. London: Centre for Economic Policy Research.
Armstrong, A., and A. Silberston (1965). Size of plant, size of enterprise and concen-
tration in British manufacturing industry 1935–1958. Journal of the Royal Statistical
Society, series A, 128: 395–420.
Asch, P., and J. J. Seneca (1976). Is collusion profitable? Review of Economics and Sta-
tistics, 58: 1–12.
Bound, J., C. Cummins, Z. Griliches, B. H. Hall, and A. Jaffe (1984). Who does R&D
and who patents? In Z. Griliches, ed., R&D, Patents, and Productivity. Chicago: Uni-
versity of Chicago Press.
Bowley, A. L. (1924). The Mathematical Groundwork of Economics. Oxford: Oxford Uni-
versity Press.
Brander, J. A., and B. J. Spencer (1985). Tacit collusion, free entry and welfare. Journal
of Industrial Economics, 33: 277–294.
Bresnahan, T. F., and P. C. Reiss (1991). Entry and competition in concentrated markets.
Journal of Political Economy, 99: 977–1009.
Broadberry, S. N., and N. F. R. Crafts (1996). British economic policy and industrial
performance in the early post-war years. Business History, 38: 65–91.
Broadberry, S. N., and N. F. R. Crafts (2000). Competition and Innovation in 1950’s
Britain. Working Paper in Economic History no. 57. London School of Economics.
Brock, W. A., and J. A. Scheinkman (1985). Price setting supergames with capacity
constraints. Review of Economic Studies, 52: 371–382.
Burn, D., ed. (1958). The Structure of British Industry. 2 vols. Cambridge: Cambridge
University Press.
Cameron, A. C., and P. K. Trivedi (1998). Regression Analysis of Count Data. Cambridge:
Cambridge University Press.
Caves, R. E. (1988). Trade exposure and changing structures of US manufacturing
industries. In A. M. Spence and H. A. Hazard, eds., International Competitiveness.
Cambridge, MA: Ballinger.
Chevalier, J. A. (1995). Capital structure and product market competition: Empirical
evidence from the supermarket industry. American Economic Review, 85: 415–435.
Cohen, W. (1995). Empirical studies of innovative activity. In P. Stoneman, ed., Hand-
book of the Economics of Innovation and Technological Change. Oxford: Blackwell.
Conyon, M., and S. Machin (1991). Market structure and the empirical specification of
price-cost margins. Economics Letters, 35: 227–231.
Corley, T. A. B. (1966). Domestic Electrical Appliances. London: Jonathan Cape.
References 513
Cowling, K., and M. Waterson (1976). Price-cost margins and market structure. Eco-
nomica, 43: 267–274.
Crepon, B., and E. Duguet (1997). Research and development, competition and inno-
vation: Pseudo-maximum likelihood and simulated maximum likelihood methods
applied to count data models with heterogeneity. Journal of Econometrics, 79: 355–378.
Curry, B., and K. D. George (1983). Industrial concentration: A survey. Journal of
Industrial Economics, 31: 203–255.
Cyert, R. M., and M. H. deGroot (1973). An analysis of cooperation and learning in a
duopoly context. American Economic Review, 63: 24–37.
d’Aspermont, C., and M. Motta (2000). Tougher price competition or lower concen-
tration: A trade-off for antitrust authorities? In G. Norman and J.-F. Thisse, eds., Market
Structure and Competition Policy: Game-Theoretic Approaches. Cambridge: Cambridge
University Press.
Dasgupta, P., and J. Stiglitz (1980). Industrial structure and the nature of innovative
activity. Economic Journal, 90: 266–293.
Davidson, C., and R. Deneckere (1990). Excess capacity and collusion. International
Economic Review, 31: 521–541.
Davies, S. (1980). Minimum efficient size and seller concentration: An empirical prob-
lem. Journal of Industrial Economics, 28: 287–301.
deMelo, J., and S. Urata (1986). The influence of increased foreign competition on
industrial concentration and profitability. International Journal of Industrial Organi-
zation, 4: 287–304.
Demsetz, H. (1973). Industry structure, market rivalry, and public policy. Journal of
Law and Economics, 16: 1–9.
Dick, A. R. (1996b). When are cartels stable contracts? Journal of Law and Economics, 39:
241–283.
Director General of Fair Trading (DGFT) (1981). Annual Report of the Director General of
Fair Trading for the Period January 1980 to December 1980. London: H.M.S.O.
Dixit, A. (1979). A model of duopoly suggesting a theory of entry barriers. Bell Journal
of Economics, 10: 20–32.
Dixit, A., and J. Stiglitz (1977). Monopolistic competition and optimal product diver-
sity. American Economic Review, 67: 297–308.
Edgeworth, F. Y. (1881). Mathematical Psychics: An Essay on the Application of Mathe-
matics to the Moral Sciences. London: Kegan Paul.
Eissa, N. (1995). Taxation and Labour Supply of Married Women: The Tax Reform Act
of 1986 as a Natural Experiment. NBER Working Paper no. 5023. Cambridge, MA:
National Bureau of Economic Research.
Eissa, N., and J. B. Liebman (1996). Labour supply response to the earned income tax
credit. Quarterly Journal of Economics, 111: 605–637.
514 References
Elliott, D. C., and J. D. Gribbin (1977). The abolition of cartels and structural change in
the United Kingdom. In A. P. Jacquemin and H. W. deJong, eds., Welfare Aspects of
Industrial Markets. Leiden: Martinus Nijhoff.
Evely, R., and I. M. D. Little (1960). Concentration in British Industry: An Empirical Study
of the Structure of Industrial Production, 1935–51. Cambridge: Cambridge University
Press.
Farber, S. (1981). Buyer market structure and R&D effort: A simultaneous-equations
model. Review of Economics and Statistics, 63: 336–345.
Fershtman, C., and N. Gandal (1994). Disadvantageous semicollusion. International
Journal of Industrial Organisation, 12: 141–154.
Fershtman, C., and E. Muller (1986). Capital investments and price agreements in
semicollusive markets. RAND Journal of Economics, 17: 214–226.
Fershtman, C., and A. Pakes (2000). A dynamic oligopoly with collusion and price
wars. RAND Journal of Economics, 21: 207–236.
Fraas, A. G., and D. F. Greer (1977). Market structure and price collusion: An empirical
analysis. Journal of Industrial Economics, 26: 21–44.
Freyer, T. (1992). Regulating Big Business: Antitrust in Great Britain and America, 1880–
1990. Cambridge: Cambridge University Press.
Friedman, J. W., and J.-F. Thisse (1993). Partial collusion fosters minimum product
differentiation. RAND Journal of Economics, 24: 631–645.
Friedman, J. W., and J.-F. Thisse (1994). Sustainable collusion in oligopoly with free
entry. European Economic Review, 38: 271–283.
Fudenberg, D., and J. Tirole (1991). Game Theory. Cambridge, MA: M.I.T. Press.
Gasmi, F., J. J. Laffont, and Q. Vuong (1992). Econometric analysis of collusive behavior
in a soft-drink market. Journal of Economics and Management Strategy, 1: 277–312.
Geroski, P. A. (1990). Innovation, technological opportunity, and market structure.
Oxford Economic Papers, 42: 586–602.
Geroski, P. A. (1991). Market Dynamics and Entry. Oxford: Basil Blackwell.
Geroski, P. A. (1994). Market Structure, Corporate Performance and Innovative Activity.
Oxford: Clarendon Press.
Geroski, P. A., and R. Pomroy (1990). Innovation and the evolution of market struc-
ture. Journal of Industrial Economics, 38: 299–314.
Guenault, P. H., and J. M. Jackson (1974). The Control of Monopoly in the United Kingdom.
2nd ed. London: Longman.
Häckner, J. (1994). Collusive pricing in markets for vertically differentiated products.
International Journal of Industrial Organisation, 12: 155–177.
Hannah, L. (1979). Mergers, cartels and concentration: Legal factors in the US
and European experience. In N. Horn and J. Kocka, eds., Law and the Formation of the
Big Enterprises in the 19th and Early 20th Centuries. Göttingen: Vandenhoeck &
Ruprecht.
References 515
Hannah, L., and J. A. Kay (1977). Concentration in Modern Industry: Theory, Measurement
and the UK Experience. London: Macmillan.
Harrington, J. E. (1989). Collusion and predation under (almost) free entry. Interna-
tional Journal of Industrial Organization, 7: 381–401.
Harrington, J. E. (1991). The joint profit maximum as a free-entry equilibrium outcome.
European Economic Review, 35: 1087–1101.
Hart, P. E. (1985). Recent Trends in Concentration in British Industry. NIESR Discussion
Paper no. 82. London: National Institute of Economic and Social Research.
Hart, P. E., and R. Clarke (1980). Concentration in British Industry 1935–75. Cambridge:
Cambridge University Press.
Hart, P. E., and E. Morgan (1977). Market structure and economic performance in the
United Kingdom. Journal of Industrial Economics, 25: 177–193.
Hart, P. E., and S. J. Prais (1956). The analysis of business concentration: A statistical
approach. Journal of the Royal Statistical Society, series A, 119: 150–191.
Hart, P. E., M. A. Utton, and G. Walshe (1973). Mergers and Concentration in British
Industry. Cambridge: Cambridge University Press.
Hatch, J. H. (1970). Competition in the British White Goods Industry. Ph.D. thesis, Cam-
bridge University.
Hatch, J. H. (1972). Stop-go and the domestic appliance industry: A case study. Bulletin
of the Oxford University Institute of Economics and Statistics, 34: 345–358.
Hausman, J., B. H. Hall, and Z. Griliches (1984). Econometric models for count data
with an application to the patents-R&D relationship. Econometrica, 52: 909–938.
Hay, G. A., and D. Kelley (1974). An empirical survey of price fixing conspiracies.
Journal of Law and Economics, 17: 13–38.
Heath, J. B. (1961). Restrictive practices and after. Manchester School of Economic and
Social Studies, 29: 173–202.
Heath, J. (1963). Still Not Enough Competition? Business Restrictive Practices Re-examined.
2nd ed. London: Institute of Economic Affairs.
Hsiao, C. (1986). Analysis of Panel Data. Cambridge: Cambridge University Press.
Hunter, A. (1966). Competition and the Law. London: George Allen.
Kamien, I., and N. L. Schwartz (1982). Market Structure and Innovation. Cambridge:
Cambridge University Press.
516 References
Kuipers, J. D. (1950). Resale Price Maintenance in Great Britain: With Special Reference to
the Grocery Trade. Wageningen: N. V. Drukkerij ‘‘Vada.’’
Levin, R. C., and P. C. Reiss (1988). Cost-reducing and demand-creating R&D with
spillovers. RAND Journal of Economics, 19: 538–556.
Lunn, J. (1986). An empirical analysis of process and product patenting: A simultaneous-
equation framework. Journal of Industrial Economics, 34: 319–330.
Lydall, H. F. (1958). Aspects of competition in manufacturing industry. Bulletin of the
Oxford University Institute of Statistics, 20: 319–337.
Lyons, B., and C. Matraves (1996). Industrial concentration. In S. Davies, B. Lyons, et
al., Industrial Organization in the European Union: Structure, Strategy and the Competitive
Mechanism. Oxford: Clarendon Press.
MacKinnon, J. G., and H. White (1985). Some heteroskedasticity-consistent covariance
matrix estimators with improved finite sample properties. Journal of Econometrics, 29:
305–325.
Mercer, H. (1995). Constructing a Competitive Order: The Hidden History of British Anti-
trust Policies. Cambridge: Cambridge University Press.
Mercer, H. (1998). The Abolition of Resale Price Maintenance in Britain in 1964: A Turning
Point for British Manufacturers? Working Paper in Economic History no. 39. London
School of Economics.
Meyer, B. D. (1995). Natural and quasi-experiments in economics. Journal of Business
and Economic Statistics, 13: 151–161.
Milward, A. S., and G. Brennan (1996). Britain’s Place in the World: A Historical Inquiry
into Import Controls, 1945–1960. London: Routledge.
Monopolies and Restrictive Practices Commission (1955). Collective Discrimination: A
Report on Exclusive Dealing, Collective Boycotts, Aggregated Rebates and Other Discrim-
inatory Trade Practices. London: H.M.S.O.
Monopolies and Restrictive Practices Commission (1957). Report on the Supply and
Exports of Electrical and Allied Machinery and Plant. London: H.M.S.O.
Monopolies Commission (1963). Report on the Supply of Electrical Equipment for Mechan-
ically Propelled Land Vehicles. London: H.M.S.O.
Monopolies Commission (1973). Parallel Pricing: A Report on the General Effect on the
Public Interest of the Practice of Parallel Pricing. London: H.M.S.O.
Montagna, C. (1995). Monopolistic competition with firm-specific costs. Oxford Eco-
nomic Papers, 47: 318–328.
Morgan, A. D., and D. Martin (1975). Tariff reductions and UK imports of manufac-
tures: 1955–1971. National Institute Economic Review, no. 72: 38–54.
Motta, M. (1992). Co-operative R&D and vertical product differentiation. International
Journal of Industrial Organization, 10: 643–661.
Mukhopadhaya, A. K. (1985). Technological progress and change in market concen-
tration in the U.S., 1963–1977. Southern Economic Journal, 52: 141–149.
National Board for Prices and Incomes (NBPI) (1968). Prices of Secondary Batteries.
Report no. 61. London: H.M.S.O.
National Board for Prices and Incomes (NBPI) (1971). Prices, Profits and Costs in Food
Distribution. Report no. 165. London: H.M.S.O.
Nelson, R. L. (1959). Merger Movements in American Industry, 1895–1956. Princeton, NJ:
Princeton University Press.
Nicholson, R. J., and S. Gupta (1960). Output and productivity changes in British
manufacturing industry, 1948–1954: A study from Census of Production data. Journal
of the Royal Statistical Society, series A, 123: 427–459.
Nickell, S. (1995). The Performance of Companies: The Relationship Between the External
Environment, Management Strategies and Corporate Performance. Oxford: Blackwell.
Nocke, V. (2000). Monopolisation and Industry Structure. Nuffield College Economics
Paper 2000-w27. Oxford University.
Norman, G., and J.-F. Thisse (1996). Product variety and welfare under tough and soft
pricing regimes. Economic Journal, 106: 76–91.
518 References
O’Mahony, M., and N. Oulton (1990). Industry-Level Estimates of the Capital Stock in UK
Manufacturing, 1948–85. NIESR Discussion Paper no. 172. London: National Institute
of Economic and Social Research.
Overstreet, T. R. (1983). Resale Price Maintenance: Economic Theories and Empirical Evi-
dence. Washington, DC: Federal Trade Commission, Bureau of Economics.
Pakes, A., and M. Schankerman (1984). The rate of obsolescence of patents, research
gestation lags, and the private returns to research resources. In Z. Griliches, ed., R&D,
Patents, and Productivity. Chicago: University of Chicago Press.
Pavitt, K., ed. (1980). Technical Innovation and British Economic Performance. London:
Macmillan.
Robinson, W. T., and J. Chiang (1996). Are Sutton’s predictions robust? Empirical
insights into advertising, R&D, and concentration. Journal of Industrial Economics, 44:
389–408.
Robson, M., J. Townsend, and K. Pavitt (1988). Sectoral patterns of production and use
of innovations in the UK: 1945–1983. Research Policy, 17: 1–14.
Rowley, C. K. (1966). The British Monopolies Commission. London: George Allen.
Salop, S. (1979). Monopolistic competition with outside goods. Bell Journal of Economics,
10: 141–156.
Sawyer, M. C. (1971). Concentration in British manufacturing industry. Oxford Eco-
nomic Papers, 23: 352–383.
Scherer, F. M. (1992). International High-Technology Competition. Cambridge, MA: Har-
vard University Press.
References 519
Shaked, A., and J. Sutton (1990). Multiproduct firms and market structure. RAND
Journal of Economics, 28: 605–628.
Shapley, L., and M. Shubik (1969). Price strategy oligopoly with product differen-
tiation. Kyklos, 22: 30–44.
Shaw, R. W., and C. J. Sutton (1976). Industry and Competition: Industrial Case Studies.
London and Basingstoke: Macmillan.
Shubik, M. (with R. Levitan) (1980). Market Structure and Behavior. Cambridge, MA:
Harvard University Press.
Simon, H. A., and C. P. Bonini (1958). The size distribution of business firms. American
Economic Review, 48: 607–617.
Slade, M. (1995). Product rivalry with multiple strategic weapons: An analysis of price
and advertising competition. Journal of Economics and Management Strategy, 4: 445–476.
Sleuwaegen, L., and H. Yamawaki (1988). The formation of the European common
market and changes in structure and performance. European Economic Review, 32:
1451–1475.
Sutton, J. (1996). Technology and market structure. European Economic Review, 40: 511–
530.
Townsend, J., F. Henwood, G. Thomas, K. Pavitt, and S. Wyatt (1981). Science and
Technology Indicators for the UK: Innovations in Britain Since 1945. SPRU Occasional
Paper Series no. 16. University of Sussex.
Utterbach, J. M., and F. F. Suarez (1993). Innovation, competition, and industry struc-
ture. Research Policy, 22: 1–21.
Utton, M. A. (1970). Industrial Concentration. Harmondsworth, UK: Penguin.
Vickers, J. (1989). The nature of costs and the number of firms at Cournot equilibrium.
International Journal of Industrial Organization, 7: 503–509.
References 521
Vives, X. (1985). On the efficiency of Bertrand and Cournot equilibria with product
differentiation. Journal of Economic Theory, 36: 166–175.
Blanket Manufacturers Association, 391 British Bronze and Brass Ingot Manufac-
Blundell, R., 225, 248, 255n turers Association, 359
Board industry, 408 British Carton Association, 408
Board of Trade, 22, 26, 31, 49, 53, 54 British Celanese, 387
annual reports, 48–49, 50, 335. See also British Chemical Ware Manufacturers
various industries Association, 401
report on resale price maintenance. See British Cinema and Theatre Printers
Resale price maintenance, Lloyds’ Association, 411
Committee report on British Coking Industry Association, 347
surveys of cartels, 52, 54–55 British Concrete Pipe Association, 403
Boatbuilding industry, 378 British Constructional Steelwork Associ-
Boiler and boilerhouse plant industry, ation, 367
35, 367 British Cycle and Motor Cycle Industries
Bolt, nut, screw, rivet, etc. industry, 382– Association, 379
383 British Dextrine Manufacturers Associa-
Bonini, C. P., 98n11 tion, 344
Book Cloth Association, 389 British Elastic Braid Manufacturers
Book publishing industry, 32, 410 Association, 393
Boone, J., 224 British Electrical and Allied Manufac-
Bougeon-Maasen, F., 191 turers Association, 372
Bound, J., 232 British Electrical Industry Fair Trading
Bounds approach. See Market structure, Council, 378n
bounds approach to British Electro-ceramic Manufacturers
Bowley, A. L., 101n14 Association, 400
B.P. (British Petroleum), 347, 348 British Embroidery Transfer Manufac-
Bradford and District Waste Pullers turers Association, 411
Association, 391 British Fabric Federation, 390, 392
Branded Knitting Wool Association, British Federation of Master Printers,
390 411
Brander, J. A., 100n British Firework Manufacturers Safety
Bread industry, 339–340 Association, 354
Brennan, G., 38, 39n British Fluorspar Producers Association,
Bresnahan, T. F., 93n6 338
Briar Pipe Trade Association, 414 British Furniture Manufacturers Fed-
Brick industry, 399 erated Associations, 406
Bridge and Constructional Ironwork British Furniture Trade Confederation,
Association, 367 405, 406
Bridgewater Roofing Tile Manufacturers British Hacksaw Makers Association,
Association, 399 381, 382
Bright Bolt and Nut Manufacturers British Hard Metal Association, 381
Association, 382 British Hot Finished Tube Conference,
British Acetylene Association, 348 357
British Aluminium Foil Rollers Associa- British Impulse Clock Manufacturers
tion, 359 Association, 369
British Aluminium Hollow-ware Manu- British Iron and Steel Federation, 347,
facturers Association, 384 356
British Bacon Agents Association, 340 British Ironfounders Association, 358
British Bacon Curers Federation, 340 British Jacquard Engineers Association,
British Basic Slag, 353 363
British Bath Manufacturers Association, British Knitting Machine Builders Asso-
358 ciation, 363
British Bottle Association, 288 British Laboratory Ware Association, 401
526 Index
British Lock and Latch Manufacturers British wine, cider, and perry (fermented
Association, 385 pear juice) industry, 346
British Malleable Tube Fittings Associa- British Wire Netting Manufacturers
tion, 358 Association, 383n
British Mantle Association, 396 British Wrought Iron Association, 355
British Mat and Matting Manufacturers Broadberry, S. N., 37n12, 67–69, 227,
Association, 395 275
British Metal Spectacle Manufacturers Broadcast receiving and sound reproduc-
Association, 370 ing equipment industry, 374–375
British Metal Window Association, 385 Brock, W. A., 73
British Motor Trade Association, 379 Brushes and brooms industry, 412
British Non-Ferrous Metals Federation, Brush Wood Turners Association, 407
359 Buffalo Picker Manufacturers Associa-
British Nylon Spinners, 387, 388 tion, 396
British Oil Burner Manufacturers Asso- Buffer Rod Makers Association, 368
ciation, 368 Builders’ woodwork industry, 405
British Ophthalmic Lens Manufacturers Building materials and mineral products
Association, 370 industries, miscellaneous, 402–405
British Paint Advisory Council, 351 Built-up Roofing Council, 403
British Paper and Board Makers Associa- Burn, D., 36, 48, 338, 348, 350, 361, 382,
tion, 407, 408 399, 400
British Paper Bag Federation, 408 Bus Seat Frame Association, 379
British Paper Box Federation, 408 Butter industry, 341
British Paper Machine Felt Association, Butter Makers and Packers Association,
391 341
British Pump Manufacturers Association,
361 Cable and wire (insulated) industry, 36,
British Pyrotechnists Association, 354 372–373
British Radio Equipment Manufacturers Cadmium Copper and Bronze Associa-
Association, 375 tion, 359, 383
British Radio Valve Manufacturers Asso- Cameron, A. C., 255, 255n
ciation, 374 Campbell, A., 24n
British Refrigeration Association, 365, Can and metal box industry, 384
376 Can Manufacturers Association, 384
British Sanitary Earthenware Manufac- Canvas Hose Manufacturers Association,
turers Association, 400 389
British Sanitary Fireclay Association, Canvas sack and bag industry, 394
399 Capital intensity
British Starter Battery Association, 377 and collusion, 78, 79, 80, 83, 128
British Sulphate of Ammonia Federation, and concentration, 79, 128
353 measures of, 75–76
British Sulphate of Copper Association, Capital stock data, 75–76, 116–117, 120n,
348 300–301
British Surgical Trades Association, 369 Car and commercial vehicle industry,
British Synchronous Clock Conference, 379
369 Carpet industry, 393
British Tanning Extract Manufacturers Cartel policy, 1–2, 8, 12–13, 22–23, 24n,
Association, 352 33n8, 42, 146, 324, 328. See also Com-
British Tarpaviers Federation, 404 petition policy; Price competition;
British Teapot Manufacturers Associa- Restrictive Trade Practices Act (1956)
tion, 400 and industrial structure in the UK and
British Whiting Federation, 403 the US, 88–90, 278, 329
Index 527
Pin and Allied Trades Association, 386 and advertising, 15–16, 19, 149–150,
Plaited Cordage Association, 392 151–172, 178–179, 199–202, 205–211,
Plant numbers 221–222, 276, 328, 329–330
and market size. See Market size, and and concentration, 14–16, 18, 87–90,
firm/plant numbers 91–104, 127–129, 138–147, 149–150,
and price competition. See Price compe- 151–172, 178–179, 185–187, 188–198,
tition, and firm/plant numbers 208–211, 221–222, 228–229, 229–238,
constructing the samples for the analy- 261–264, 265–274, 274–275, 283, 327–
sis of, 294–297, 303–304 328, 329, 330, 331
data, 76, 117–119, 120n, 297 and firm/plant numbers, 17, 277–279,
econometric model for, 302–303, 309– 280, 283, 284–285, 304–309, 312–313,
312 314–315, 318, 319, 323–325, 327–328,
econometric results for, 312–313, 314– 330
315, 318, 319, 322–323 and innovation, 16, 19, 228–229, 229–
Plaster Ventilators Manufacturers Asso- 238, 252–253, 257–260, 274–276, 328,
ciation, 405 330, 331
Plastic products industry, 413 and market structure, 3. See also Price
Plastics Hardware Association, 413 competition, and concentration; Price
Plastic Spectacle Manufacturers Associa- competition, and firm/plant numbers
tion, 370 and profitability/profits, 3, 14, 15, 17–
Plate Glass Association, 401 18, 87, 146, 147, 277–279, 280–285,
Plate glass industry, 400–401 304–309, 313, 316–322, 323–325, 327–
Plywood and textured board industry, 328, 329
405 Price-fixing, 22, 23, 27, 28, 34, 35, 53, 57,
Pneumatic Metal Working Tool Associa- 58, 60, 61, 333, 334, 335. See also various
tion, 366 industries
Pneumatic Tool Association, 366 Price, R., 301
Polishes industry, 353 Prices and incomes policy, 42, 184, 309
Political and Economic Planning, 38, Primary battery industry, 376
39n Printing, bookbinding, and paper goods
survey of trade associations, 33n9, 38, making machinery industry, 365
48, 49–50, 60, 81, 335. See also various Printing (general) industry, 411
industries Printing ink industry, 354
Pomroy, R., 226n3 Procter and Gamble, 351
Portable Air Compressor Association, Product differentiation
362 horizontal, 94–95, 95n7, 101, 124, 169,
Portable power tool industry, 366 180, 230n6, 251
Post Office, 373, 374 vertical, 73, 82, 83, 95n7, 150, 164. See
Potassium Carbonate Association, 348 also Quality
Pot Still Malt Distillers Association, 345 Productivity. See Efficiency
Powered industrial truck and industrial Profitability, 3, 17–18, 19–20, 324, 326, 330
tractor industry, 364 and market structure as jointly endoge-
Prais, S. J., 98n11 nous variables. See Market structure,
Pratten, C. F., 122 and profitability/profits as jointly
Precast concrete goods industry, 402–403 endogenous variables
Precision File Association, 381 and price competition. See Price compe-
Precision Winding Association, 388 tition, and profitability/profits
Pressed Bowl Makers Association, 369 constructing the samples for the analy-
Pressed Brick Makers Association, 399 sis of, 294–297, 303–304
Pressed Steel, 217 data, 297–298
Price competition. See also Competition; econometric model for, 302–303, 309–
Intensity of price competition 312
Index 537